ClimateChangeFork

Figure 1 – Almost three-quarters of the global population live in countries that are net importers of fossil fuels – Fossil fuel exporters – Wikipedia

Last week’s blog started to discuss attempts to shift the energy transition to take sovereignty into account. The point was made that such shifts will be compatible with other objectives such as security and the mitigation of climate change. The top figure of this blog shows that about 75% of the global population lives in countries that cannot make such a shift based on fossil fuels. They don’t have enough of these resources on hand and need to import them. Even some of the countries that have enough fossil fuels and can export them have issues. Many of the exporters of fossil fuel are in the Middle East and are directly impacted by the Iran war with US and Israel, which has made the Strait of Hormuz almost impossible to pass exports through (see March 25, 2026 blog). Russia is one of the world’s top fossil fuel exporters. However, Russia’s attack on Ukraine triggered an economic boycott by many Western countries. Figure 2 shows the extent of that move.

Figure 2 (Source: Forbes)

The US is marked in the top figure as a net exporter of fossil fuels, an “honor” that it acquired in 2019 with the advance of fracking and its ability to find new resources. However, prices can play an important role, and net exporting countries can have a significant import. The United States is probably the most consequential of these from a global perspective. Figure 3 describes oil imports into the US in 2021. Two years later, it was declared a net exporter.

Figure 3 – While the US is the world’s largest oil producer, it still imports more than 40% of the oil and petroleum products it consumes (Source: Visual Capitalist)

At the time Figure 3 was created, US import of fossil fuels was still about 40%. Presently it has been reduced to 17%. As Figure 3 shows, the largest source of imported oil was Canada. However, our current relationship with Canada is not the best! All these global events have direct impacts on energy prices. I will cite below two cases of net importers of fossil fuels: Western Europe with an emphasis on Germany—and Pakistan. Next week’s blog will generalize the impacts to more countries and will try to explore impacts on energy efficiency.

Western Europe:

The wholesale price for European natural gas soared in June to 60 euros per megawatt-hour, the highest it has been since the war with Iran began, and has continued to hover around that level. Without a lasting resolution to the war, analysts said, Europe will be forced to keep paying higher prices to fill its natural gas storage sites, costs that will trickle down to utility bills.

Germany:

New data from London think-tank Energy Institute found that together wind and solar generated 44 per cent of the country’s electricity in 2025 – overtaking fossil fuels by one per cent. Since the introduction of its landmark renewable energy law (Erneuerbare-Energien-Gesetz) in 2000, the country’s share of generation from wind and solar alone has skyrocketed by around 42 per cent. At the same time, coal – which is often described as the ‘dirtiest’ form of energy – fell from supplying more than half of Germany’s electricity to just 21 per cent. Experts are now hopeful that the country will wean itself off coal before 2038. It’s an impressive feat considering Germany phased out nuclear power – which contributed to 6.6 per cent of the country’s total power production in 2022. While nuclear generation is often categorised as clean energy, concern around its harmful waste and environmental impacts remains rife. It mirrors a broader trend, as data from energy think-tank Ember found that wind and solar generated more power than fossil fuels across the EU for the first time last year (30 per cent compared to 29 per cent). Despite the victory, Germany is still grappling with negative electricity prices – which occur when supply outstrips demand. This is often attributed to the inflexible nature of wind and solar power, which generate electricity based on weather conditions rather than actual demand. It has led to a spike in curtailment, where operators temporarily switch off or reduce the output of solar and wind farms. When curtailment occurs directly because generating electricity is no longer economically viable (rather than because of grid constraints), this is known as price-sensitive, or commercial, curtailment.

Pakistan:

Electric motorcycles are a crucial indicator of changes driven by war. In poorer countries, where people are most vulnerable to energy-price shocks, two-wheelers are far more common than cars. Electric models typically cost more to buy than gasoline-powered versions.

But rising gasoline prices have given electrics a big advantage: It’s cheaper to charge a battery than to fill up a fuel tank. In Vietnam, famous for its motorcycle-swarmed streets, the biggest domestic producer, VinFast, has seen soaring sales. In Kenya, about 25 to 35 percent of new motorcycle sales are electric, said Gagan Gupta, the founder of Spiro, the largest electric vehicle company operating in Africa. He said the energy shock helped “multiply everything by two times.” In all these places, people in the industry increasingly tout the idea that future generations will never operate a gasoline-powered vehicle. “The demand is effectively open-ended,” said Josh Whale, the founder of Ampersand, an electric-motorcycle company that operates in Kenya and Rwanda.

However, sovereignty is not the only shift to local control that the energy system is trying to adjust to. Other decentralization attempts are in play too (Future of Energy: Embracing Decentralization):

Welcome to the world of decentralization and energy—a dynamic duo shaping the future of how we power our lives. Decentralization is like giving power back to the people, literally. Instead of relying on big, centralized power plants, it’s all about spreading energy generation and control to local sources. Traditionally, we’ve been hooked on massive power plants and extensive networks to meet our global energy needs. But decentralization flips the script, advocating for a more local and diverse approach to energy production and management. This shift promises resilience, sustainability, and even a touch of democracy in how we handle our energy resources. In this exploration, we’ll unravel the simplicity and potential behind decentralization and energy. Get ready to discover how this shift is not just changing the way we power our world but also paving the way for a brighter, greener, and more accessible energy future.

We need more time to quantify these efforts.

Figure 1 – Average price of a solar panel vs. global electricity generated from solar, 1975-2025 (Source: Pew Research Center)

Figure 1 was taken from the Pew Research Center site that was discussed in last week’s blog (August 6, 2026), which ended with the following lines: “There is a growing realization that the global energy transition to mitigate climate change coincides with energy security, but both need local control (IEA’s Latest Report Reveals Energy’s Shift From Climate to Security). However, neither affordability nor local control are assured.” This blog goes through some details about recent attempts to include access to energy as part of state sovereignty. I will start with the Merriam-Webster definition of sovereignty and its present status:

1 a : supreme power especially over a body politic

b : freedom from external control : autonomy sense 1

c : controlling influence

2 : one that is sovereign

especially : an autonomous state : autonomy sense 2

Number of sovereign countries (AI Through Google, now Gemini):

There are 195 widely recognized sovereign countries in the world. This total includes the 193 member states of the United Nations, plus two permanent observer states: the Holy See (Vatican City) and the State of Palestine. [1, 2]

Depending on the criteria used for political recognition, some lists count up to 197 by including associated states like the Cook Islands and Niue, or higher if counting partially recognized or disputed territories. [1, 2, 3]

Figures 2 and 3 illustrate attempts to integrate energy affordability with the security of sovereign countries through local control.

Figure 2 (Source: Canary Media)

Figure 3 illustrates that China is the facilitator of this transition:

Figure 3

All three figures illustrate that the transition to local control of energy sources ties in well with the transition from coal to solar energy. However, in spite of the fact that China is leading that transition, the country is still responsible for more than 50% of global use of coal—the dirtiest fossil fuel that underlies recent energy use (Ranked: The Countries That Consume the Most Coal):

Figure 4 – The countries that consume most coal

Trump claimed security as the reason behind his attempt to shift legislative issues to executive control under his “Liberation Day” global tariff policy. Three paragraphs from that publication give insight into the Trump administration’s thinking:

On April 2, 2025—a date President Trump proclaimed “Liberation Day”—the administration announced the most sweeping tariff hike since the Smoot-Hawley Tariff Act, the 1930 law best remembered for triggering a global trade war and deepening the Great Depression.

At the center of the Trump administration’s trade doctrine lies a conviction that the U.S. trade deficit with each partner is not just a data point, but a ledger of national failure. It is a scarlet letter testifying to decades of asymmetric trade agreements, in which the United States bartered away its industrial birthright for mere corporate margins and cheaper imports.

What the United States gained in shareholder returns, it lost in productive capacity. Steel, aluminum, and shipbuilding were left to wither—casualties of an indifferent trade regime. And this, in Trump’s framing, is where trade policy crosses the Rubicon. What began as economic decline becomes a national security liability, one that would be exposed in the event of a sustained military conflict. After all, it was the United States’ unmatched industrial base, not just its battlefield tactics, that secured victory in World War II. Today, the argument goes, that the “arsenal of democracy” has been allowed to rust, while China has assembled the very industrial capacity that once defined U.S. strength. 

However, the shift to local control of the economy is not restricted to the US. In many cases, it is driven by smaller countries’ interests in breaking their dependence on the US. My next blog will list some such efforts in countries that are ahead in the transition to local control.

 Figure 1 – AI-generated (Gemini) image response to a request for a picture of different ways for an energy transition

The picture above does not mention climate change or any other trigger. Even so, the picture illustrates the full transition that would serve to mitigate climate change. In the US, however, the Trump administration is advancing policies to ignore climate change and stop the energy transition:

When the Trump administration declared two weeks ago that it would largely disregard the economic cost of climate change as it sets policies and regulations, it was just the latest step in a multipronged effort to erase global warming from the American agenda.

But President Trump is doing more than just turning a blind eye to the fact that the planet is growing hotter. He is weakening the country’s capacity to understand global warming and to prepare for its consequences.

The administration has dismantled climate research, firing some of the nation’s top scientists, and gutted efforts to chart how fast greenhouse gases are building up in the atmosphere and what that means for the economy, employment, agriculture, health and other aspects of American society. The government will no longer track major sources of greenhouse gases, data that has been used to measure the scale and identify sources of the problem for the past 15 years.

Still, anthropogenic global climate change continues to accelerate:

Global warming appears to have shifted into a faster gear. After removing the effects of El Niño, volcanic eruptions, and solar cycles, researchers found that the planet has warmed at roughly 0.35°C per decade over the past ten years, compared with just under 0.2°C per decade from 1970 to 2015. The acceleration, which becomes visible around 2013 to 2015, appeared across five major global temperature datasets with more than 98% statistical certainty.

Presently, not including climate change, the global events with the largest direct impacts on global energy use are the US-Israel-Iran war, the Russia-Ukraine war, and the global energy transition via electrification. Some of my earlier blogs described the impacts of the US-Israel-Iran war on the transport of fossil fuels (March 25, 2026) and the role of data centers in global electrification (May 21, 2026).

Another issue that drives countries to change energy suppliers is the desire for control. Cuba serves as a good example of what can happen to a country that loses control over its energy supply:

Since January 2026, the United States is imposing a fuel blockade on Cuba, causing an ongoing fuel shortage and economic crisis in the latter.^{[14][15][16][17]} As an island, Cuba heavily depends on importing petroleum to meet its energy needs, particularly from nearby Venezuela and Mexico. Following the 2026 United States intervention in Venezuela, in which U.S. forces ousted president Nicolás Maduro, Venezuelan oil exports to Cuba were cut off, leaving the country without adequate supply. In addition, the U.S. enacted a blockade on oil coming into Venezuela a month prior.^[18] The United States is reportedly motivated by a desire for regime change in Cuba by the end of 2026.^[19] 

In all these cases, the solution seems to be the same: get control over your energy supply through internal means. If possible, the word “sovereignty” should apply to your energy supply. The scenario shown in the top figure describes such a transition. It mitigates climate change without referring to climate change. The quote below from Bloomberg summarizes a similar situation:

(Bloomberg) — The world’s hunger for electricity, as demand rises almost everywhere, will lead to solar becoming the dominant source of power early next decade, according to new analysis from BloombergNEF. Surging demand from data centers, population growth, rising incomes and more battery-powered vehicles are all boosting demand for electricity, researchers wrote in BNEF’s annual New Energy Outlook on Tuesday. The report concludes that rapid electrification is set to accelerate the transition to clean energy. “So far this decade, the world has suffered three substantial shocks to the energy system: the Covid-19 pandemic, the war in Ukraine and, most recently, the conflict in the Persian Gulf,” the analysts wrote. “Successive energy market shocks could be a boon for the energy transition as some countries look to decouple from imported fossil fuels and bolster their energy security.”

Recently, a Pew Research Center publication summarized the global energy transition over the past 25 years. Figures 2 and 3 illustrate these trends:

Figure 2

Figure 3

The global trend of converting as much energy use as possible to electricity is illustrated by the attempt to electrify aviation, an application unthinkable until recently:

TAXIING TOWARDS the runway, the propeller stops turning as your correspondent pulls the throttle back to await take-off. On a single-engine light aircraft, that is usually a worrying sign of engine failure. Yet this plane is powered by an electric motor and, just as when you come to a halt in a Tesla, the motor stops until it is needed again. On advancing the throttle the propeller springs back to life and the aircraft is soon briskly climbing, in an eerily quiet manner. The two-seater Pipistrel Velis Electro is the first electric aircraft to be certified for pilot training. Besides its low noise, it has other unusual features. Instead of the lengthy series of engine checks a combustion-engine plane requires before take-off, the Electro is ready to go with the flick of four switches. Since it has few moving parts, it is also cheaper to maintain. And without directly producing any nasty emissions it can claim to be green, especially if recharged with renewable power. Topping up the battery takes just over an hour.

Statista illustrates the progress of global electrification with Figure 4 and the paragraph below:

Ten years after the Paris Agreement was opened for signature by member states at the UN Headquarters in New York on April 22, 2016, the global energy transition is beginning to reshape the electricity mix. As data from Ember’s Global Electricity Review 2026 shows, renewables have now overtaken coal as the world’s largest source of electricity generation. In 2025, renewables accounted for 33.8 percent of global power output, slightly ahead of coal at 33.0 percent — a symbolic milestone in the shift toward low‑carbon energy.

The trend reflects a steady but uneven rise of renewables over the past twenty‑five years, from less than 19 percent in 2000 to nearly 34 percent in 2025. Growth was relatively modest in the early 2000s, with the share hovering below 20 percent until 2010, then accelerating markedly over the past decade as wind and especially solar power expanded rapidly. In 2025, this momentum strengthened further, with the entire increase in global electricity demand met by clean energy sources. Solar power alone accounted for 75 percent of the net increase, adding 636 TWh of generation during the year.

Over the same period, the role of fossil fuels in electricity generation has followed a two-phase trajectory. Between 2000 and the mid-2010s, their combined share initially edged up, peaking at around 69 percent in 2012, as growing global demand was largely met by coal and gas. Since then, however, the trend has reversed. Coal’s share has seen the most pronounced decline, falling from about 41 percent at its peak to 33 percent in 2025, while gas and other fossil fuels have decreased more gradually from roughly 28 percent to 24.4 percent. Despite this progress, fossil fuels still account for a majority of global electricity generation (around 57 percent in 2025), highlighting the scale of the challenge ahead in meeting climate targets.

Figure 4 (source: Statista)

The interesting development in the energy transition’s shift to sovereign states is that while it is not always anchored in developed countries, it can shift sovereignty away from rich countries:

Several ambitious international initiatives that aim to deliver access to clean, modern energy services to underserved populations in developing countries have recently taken root, including the UN Sustainable Energy for All initiative, the Energy+ Partnership, and Power Africa. The scale of the challenge is great: today, 1.3 billion people lack access to electricity, yet extending the electrical grid is not a cost-effective solution in many rural, low-income areas. The International Energy Agency estimates that to achieve universal access to electricity by 2030, more than 50 percent of new electricity connections will be through distributed energy solutions, and more than 90 percent of those will use renewable energy sources. To meet the scale of the challenge, policy and regulation will need to be designed in ways that recognize the potential of distributed renewable energy to contribute to energy access goals, especially given recent technological and business model innovations of enterprises delivering these solutions. International initiatives investing in clean energy access can work with national governments and local actors to inform policy and regulation that supports distributed renewable energy enterprises. This Issue Brief collects the perspectives of some of these enterprises on policy and regulatory barriers and opportunities to scaling up clean energy access, and synthesizes these perspectives into two recommendations:

There is a growing realization that the global energy transition to mitigate climate change coincides with energy security, but both need local control. However, neither affordability nor local control is assured.

AI (Copilot) illustration of how AI should create guardrails for itself

This is the last blog I have planned in this series focusing on AI. I look at the dangers of relying on AI for decision making and some of the protective guardrails that are under discussion. The top picture was drafted by AI (through Copilot) as a response to a request to draw a picture of how AI should create guardrails for itself. The fact that there are no explanations for the various symbols, can serve as a message that AI is still “thinking” about it and also gives readers an opportunity for their own input about the question.

Ideally, AI would be used to make the world a better place for humanity. This line of thinking is becoming more popular, and it’s easy to picture this idealized future:

At its core is the concept of sufficiency – the idea that people can enjoy a prosperous, healthy life without constantly striving to consume or accumulate more material possessions that degrade the natural world on which all life depends.

To achieve this, the authors envisage three steps: more than halving average working time from 2,100 hours a year to 1,000 hours, roughly equivalent to a two-and-a-half-day working week; encouraging people to eat less red meat, which is the main driver of deforestation and ecological destruction; and refocusing the economy toward low-consumption activities by more than doubling education spending to €8,400 (£7,250) a person and healthcare spending to €14,400.

Piketty said: “One extra euro of GDP in education and health has three to four times less material footprint and energy consumption than one extra euro of GDP in the manufacturing sector. So that’s why the sectoral shifts are hugely important.”

This picture is not too different from the one Elon Musk advocated, which he called “sustainable abundance” (see my March 11, 2026 blog).  How AI can actually help in achieving a better world is a much more difficult issue.

A recent Economist interview with Elon Musk incorporated AI in trying to answer this question:

ELON MUSK, in his telling, is a modern-day Cassandra—better than most at predicting the future, but doomed to be disbelieved. “People don’t realise that what I’m saying will come to pass,” he tells The Economist in a 90-minute interview for The Insider, held at a giant Tesla factory in Texas where electric cars drive themselves off the assembly line. That may be because his future-gazing ranges from the unsettling to the downright weird. Mr Musk predicts that, within five years, artificial-intelligence systems may surpass the sum of all human intelligence. Within ten years, robots in the workplace will help usher in an era of such “amazing abundance” that money will become meaningless. And within 20 years, Britain—a country that he has not visited for years—will be engulfed in civil war.

However, most people are not interested in making the world a better place. Most are interested in making their world a better place, with more power and resources. In trying to achieve that, AI can evolve into a dangerous tool that needs guardrails. Who is setting up the guardrails is now part of the fight for power. So, I asked AI the following question:

How can AI become dangerous?

AI becomes dangerous when its capabilities, incentives, or misuse outpace human control. The most serious risks identified by hundreds of AI experts fall into five major danger zones: dangerous capabilities, weapons & cyberattacks, competitive pressures, power centralization, and false information.

⚠️ The Core Ways AI Can Become Dangerous

1. Misuse by Humans

People can deliberately use AI for harmful purposes:

• Designing bioweapons or chemical agents
• Launching cyberattacks at unprecedented scale
• Creating autonomous weapons
• Running adversarial attacks on critical infrastructure

These are not hypothetical — experts warn these risks already show measurable probability of catastrophic outcomes.

2. Misalignment with Human Goals

AI can behave in ways developers never intended:

• Reward hacking / specification gaming (finding loopholes in instructions)
• Scheming or power‑seeking behavior in pursuit of long-term goals
• Autonomous replication or actions outside human oversight

Misalignment becomes dangerous when AI systems gain:

• High autonomy
• Strategic planning ability
• Persuasive or manipulative communication skills

3. Systemic Risks to Society

AI can destabilize social and political systems:

• Concentrating power in governments or corporations
• Accelerating economic inequality
• Undermining human agency through overdependence
• Locking in current values, limiting future moral progress

These risks grow gradually but can become irreversible.

Earlier this month, the issue of AI developing into something existentially dangerous moved beyond a theoretical question and became real:

IT SOUNDS like something out of a sci-fi film. On July 16th Hugging Face, an artificial-intelligence platform, announced it had notified law enforcement after an attacker had used an AI agent to breach its systems. On July 21st it emerged that there had been no human involved; the AI agent was the attacker.

OpenAI applies safeguards to its publicly available models, which ensure they do not pursue unwanted actions. Ask the commercially available version of ChatGPT to help you hack into a website, for example, and it will refuse. To assess the capabilities of its unreleased model, the company had temporarily suspended those restrictions. Simply letting a highly capable cyberattacker loose on the open internet would have been too risky, however, so OpenAI had placed its models in a sandbox—an isolated computer environment with no internet access, except for an internally hosted third-party service that allowed it to fetch small software packages needed to complete its tests.

This author is right that it is a scenario straight out of science fiction. Just go to Google or your favorite search engine (or AI) and ask for the “top 10 robot uprising movies.”

One of my reasons for ending my series on AI and returning to the real world was finding out how prolific AI books have become in publishing:

Two business school professors, curious about A.I. books and whether anyone actually likes them, gathered data about 10 million books published on Amazon over the last five years. They found that the number of e-books published per month had tripled since the release of ChatGPT, to more than 300,000 at the end of last year, from around 100,000 in 2022. (Amazon said that its internal metrics did not show that level of growth, but would not share its figures.)

It is becoming increasingly difficult to distinguish what is real and what is not. So, I am going to return to focusing on the “real” world. Though I may occasionally use AI for some technical jobs, I will always mark the real (original) source.

In last week’s blog, I started demonstrating the power of AI to predict aspects of the future by having it analyze the future of climate change. My focus was not to demonstrate the benefits or the accuracy of these predictions. To be honest, with my background in the topic, I did not find that the predictions added anything to my knowledge base. Nor would they help teachers that teach the topic. However, the predictions might help students and people with a minimal background in the subject. The AI that I used was Gemini, which Google has now adopted for all inquiries. Two main aspects illustrate the predictions: the top picture and the text that follows. My original question was “Can you predict the future of global climate change and explain how you arrived at these predictions?” The picture in last week’s blog came out on the first try; the text needed some feedback. My prompt from last week’s blog did not ask to emphasize humanity’s role in the process, bu it came up anyway because of AI’s exposure to IPCC training. The noticeable result in last week’s blog was that the opening picture was OK on the first trial, but the text eliminated most traces of quantitative predictions in the form of numbers to focus instead on trends. This blog is using a different AI, Copilot, which is offered as part of the Microsoft system. A short description of the two AI systems is given below (through Gemini):

Gemini is Google’s multimodal AI assistant, designed for deep web research, creative ideation, and processing vast amounts of data. Copilot is Microsoft’s productivity-focused AI, built natively into the Microsoft 365 ecosystem to streamline writing, analyze spreadsheets, and summarize meetings

In this blog, I explicitly asked Copilot to include human contributions to climate change. Unlike the previous blog, the top picture needed some fine tuning to limit the amount of information included. However, the text needed none. I started the discussion with the following question:

Can you predict the future of global climate change and explain how you arrived at these predictions? Please include human contributions.

Big picture: where the climate is headed

Global climate change will continue to intensify throughout this century. The exact amount depends on how quickly humanity cuts greenhouse gas emissions, but some patterns are very robust:

• Global temperature: Under all scenarios, global surface temperature keeps rising at least until mid‑century. Without deep emission cuts, warming is likely to exceed and then above pre‑industrial levels this century.
• Extremes and weather: More frequent and intense heatwaves, heavy rainfall, droughts in some regions, and a higher proportion of intense tropical cyclones are expected as warming increases.
• Ice, oceans, and sea level: Continued warming drives loss of Arctic sea ice, glaciers, and ice sheets, long‑term sea‑level rise, and more marine heatwaves and ocean acidification. Many of these changes are irreversible for centuries to millennia.
• Water cycle: The global water cycle becomes more intense, stronger monsoons, more severe wet and dry events, and greater variability in rainfall.

Human contribution: how much is “us”?

The scientific consensus is blunt:

• Human activities have “unequivocally” caused global warming. Burning fossil fuels, deforestation, and industrial processes have raised greenhouse gas concentrations and pushed global temperature to about above 1850–1900 levels (2011–2020 average).
• Anthropogenic forcing dominates recent changes. When scientists run climate models with human emissions, they reproduce the observed warming; without human emissions, the models show little warming. This is a key attribution test.
• Extremes already show our fingerprint. Many heatwaves, heavy rainfall events, and droughts now have a clear human‑caused component—meaning they are more likely, more intense, or both because of our emissions.

In short: the main driver of current and near‑future climate change is human greenhouse gas emissions, not natural variability or solar changes.

How scientists make these predictions

1. Climate models and scenarios

• Climate models: Complex computer models simulate the atmosphere, oceans, ice, and land. They are tested against past climate and observations to ensure they can reproduce known patterns and trends.
• Emissions scenarios (e.g., SSPs): Scientists feed different futures into these models—high‑emissions, medium, and strong‑mitigation pathways. Each scenario reflects assumptions about energy use, land use, population, and policy.
• Ensembles: Many models are run together to form an “ensemble,” giving a range of possible outcomes and a sense of uncertainty rather than a single number.

2. Physical understanding

• Greenhouse effect: We know from physics that increasing CO , methane, and other gases traps more heat, raising global temperature.
• Climate sensitivity: The IPCC’s best estimate is that doubling CO leads to about of warming, with a narrowed uncertainty range compared to earlier reports.

3. Attribution methods

• Detection and attribution studies: Researchers compare observed trends with model simulations that include and exclude human influences, using statistical methods to quantify how much of a change is due to us versus natural variability.

What the future looks like under different human choices

• High‑emissions path (little mitigation): Warming could reach well above and potentially or more by 2100, with severe impacts on ecosystems, food systems, health, and infrastructure, plus large and continuing sea‑level rise.
• Intermediate path (current policies / NDCs): Current national pledges make it likely that will be exceeded and harder to keep warming below unless policies are strengthened significantly and quickly.
• Strong‑mitigation path (rapid deep cuts): Rapid, sustained reductions in CO and other greenhouse gases could still limit warming close to , reduce the risk of the most extreme outcomes, and slow long‑term sea‑level rise—though some changes already locked in will continue.

Bringing it together

Direct answer: The future of global climate change is largely determined by human decisions about emissions in the next few decades. Physics, observations, and models all point to continued warming, more extremes, and long‑lasting changes to oceans and ice, with human greenhouse gases as the dominant cause.

If you’d like, I can map this into a more explicit “Venn diagram” of natural variability vs. human drivers vs. future impacts to match the diagram you asked for earlier.

I followed with one more question:

Can you quantify the predictions and the scenarios?

Here’s a quantified look at the future of global climate change based on the latest IPCC Sixth Assessment Report (AR6) and related scientific literature.

🌡️ Global Temperature Projections

Inference: Even under moderate mitigation, the world is likely to exceed 1.5°C around 2035–2040, with 2°C reached before 2060 unless emissions fall sharply.

🌊 Sea‑Level Rise

Drivers: Thermal expansion, glacier melt, and ice‑sheet loss (especially Greenland and West Antarctica).

🔥 Extreme Events

• Heatwaves: Frequency projected to double or triple by mid‑century under SSP2‑4.5.
• Heavy precipitation: Increase of 10–30% intensity in many mid‑latitude and tropical regions.
• Droughts: More severe in subtropical zones (Mediterranean, southwestern U.S., southern Africa).
• Tropical cyclones: Fewer overall, but stronger and more rainfall‑intense.

🌍 Human Contribution Quantified

• Anthropogenic greenhouse gases account for >90% of observed warming since 1850.
• CO₂ concentration: ~420 ppm (2026), up from 280 ppm pre‑industrial.
• Radiative forcing: Human activities add roughly +2.7 W/m² net forcing to Earth’s energy balance.
• Attribution studies: Human influence explains ~1.07°C of the 1°C observed warming (natural factors <0.1°C).

⚙️ How These Numbers Are Derived

1. Model ensembles (CMIP6): Dozens of global climate models simulate physical processes under standardized scenarios.
2. Observational constraints: Satellite and surface data calibrate model sensitivity.
3. Probabilistic projections: Statistical methods combine model outputs to yield ranges (e.g., 5th–95th percentile).
4. Attribution analysis: Compare simulations with and without human emissions to isolate anthropogenic effects.

🧭 Interpretation

• Near‑term (2026–2040): Warming locked in at ~1.5°C regardless of scenario.

• Mid‑century (2040–2060): Divergence begins—strong mitigation stabilizes, high‑emission accelerates.

• End‑century (2100): Range 4–4.4°C, depending on human choices.

In real life, there is an ongoing discussion about anthropogenic contributions to the heatwaves that we are currently experiencing:

The record-breaking heat that’s scorching Europe day and night this month would not have been possible without climate change, according to a new study. The World Weather Attribution rapid study released Friday found that the heat would have been virtually impossible just five decades ago and is 200 times more likely today than it would have been 20 years ago. Millions in France, Italy, Spain, the United Kingdom and elsewhere in Europe are experiencing extreme temperatures and humidity this week associated with a heat dome. Daytime temperatures have topped 40 degrees Celsius (104 Fahrenheit) in many places, while high nighttime temperatures have also made it harder to cool down and recover.

It is obvious, that on this issue, governments are not yet using AI to improve the prospects of our future. 

AI climate change predictions by Gemini

I am back to AI as a quantitative predictor (June 26, 2026 blog). I interrupted my discussion of the topic to describe events connected to Independence Day in the US on July 4^th. I will try to demonstrate the use of AI in predicting impacts of climate change. This will be a long blog. The top figure summarizes the AI prediction (with Gemini). Most of the writing included here is from my direct discussions with AI. As in previous blogs, my questions or prompts to AI are in red. The words “anthropogenic” and “attributions” are not mentioned in the questions. However, as will be seen in its predictions, AI closely follows the IPCC (Intergovernmental Panel on Climate Change) analyses, which are based on socioeconomic scenarios directly connected to the future climate, factoring in types of energy use. These concepts relate to attribution, including whether they are anthropogenic (human-caused). However, my lack of direct use of the keywords “anthropogenic” and “attributions” might (or might not) be the reason that the AI response doesn’t mention governments and societal attitude to climate change. The recent global heat waves have brought the conversation back to attribution and anthropogenic climate change.

The June 26^th blog shows that AI admitted to vulnerabilities in predictions under the conditions of a “data gap”:

The “Data Gap”: AI struggles significantly when there is a time gap between its training data and a current event. It cannot naturally account for sudden, unprecedented “black swan” events, such as geopolitical outbreaks or unrecorded financial crises. [1, 2]

• Overconfidence and Hallucination: Evaluation studies by institutions like Knowledge at Wharton show that AI models frequently exhibit overconfidence, assigning dangerously high probability scores to outcomes that lack factual evidence. [1]

• Lack of True Intuition: AI relies purely on historical data correlation. Unlike humans, it lacks genuine creative insight or “judgmental forecasting” capabilities—the ability to assign probabilities based on abstract intuition and personal domain expertise. [1, 2]

The AI predictions, to a large degree, come out of the IPCC general reports, which summarize the data on the issue in intervals of 5 – 7 years. The last one came out in 2023. One major recent change in their evaluation was to scrap the worst emissions scenario (RCP8.5). The IPCC climate reports are a good example of a “data gap,” which makes climate change predictions difficult. Let me now proceed to the AI predictions. I started my AI discussion (with Gemini) with the following question:

 “Can you predict the future of global climate change and explain how you arrived at these predictions?”

Predicting the future of global climate change involves analyzing complex, interacting systems—combining physical climate science with human socio-economic choices. Because we cannot know exactly how human society will act, scientists use scenarios rather than absolute crystal-ball predictions.

Here is what the future likely holds based on current scientific consensus, followed by the methodology used to arrive at these conclusions.

The Predictions: Where the Planet is Headed

The trajectory of climate change depends heavily on global emissions over the next few decades. Scientists generally look at three potential pathways:

1. The High-Emissions Scenario (Business as Usual)

If global reliance on fossil fuels continues to grow without significant policy shifts, global average temperatures could rise by () above pre-industrial levels by the year 2100.

• Impacts: Catastrophic sea-level rise (potentially exceeding 1 meter), widespread crop failures, frequent extreme weather events making certain equatorial regions virtually uninhabitable, and severe biodiversity loss.

2. The Moderate / Current Trajectory Scenario

Based on current global policies and climate pledges (as of the mid-2020s), the world is tracking toward a warming of roughly () by 2100.

• Impacts: While this avoids the absolute worst-case outcomes, it still triggers severe consequences. We will see heavily intensified heatwaves, prolonged droughts, destabilized ice sheets, and significant economic strain on global infrastructure, particularly in coastal and developing nations.

3. The Low-Emissions Scenario (Paris Agreement Targets)

If the world rapidly transitions to renewable energy and achieves net-zero greenhouse gas emissions by around 2050, warming could be capped between ().

• Impacts: Climate impacts will still intensify compared to today (more intense storms, shifting agricultural zones), but the changes would remain largely manageable, and society would have a much higher capacity to adapt.

How These Predictions Are Made

Climate scientists arrive at these predictions through a rigorous, multi-layered methodology that blends historical data, physics, and economic forecasting.

1. Representative Concentration Pathways (RCPs) and SSPs

Scientists don’t just guess what humans will do; they create standardized frameworks called Shared Socioeconomic Pathways (SSPs) and Representative Concentration Pathways (RCPs). These model different futures based on population growth, economic trends, energy technologies, and environmental policies. They calculate exactly how much greenhouse gas will end up in the atmosphere under each scenario.

2. Global Climate Models (GCMs)

Once a scenario is selected, the data is fed into massive supercomputers running Global Climate Models. These models divide the Earth, atmosphere, and oceans into a 3D grid. Using fundamental laws of physics, fluid dynamics, and chemistry, the supercomputers simulate how energy and mass move through the grid over decades. They calculate variables like:

• Atmospheric pressure and temperature
• Ocean currents and heat absorption
• Cloud formation and solar radiation reflection

3. Accounting for Climate Feedbacks

Predictions must account for “feedback loops,” which can accelerate or decelerate warming.

• Positive Feedback (Amplifying): As the Arctic ice melts, the darker ocean water absorbs more heat than the reflective white ice did, accelerating warming (the Albedo Effect). Similarly, thawing permafrost releases trapped methane, a potent greenhouse gas.
• Negative Feedback (Dampening): Increased carbon dioxide can sometimes stimulate increased plant growth in certain areas, which absorbs more from the atmosphere, slightly mitigating the warming effect.

4. Historical Paleoclimate Data

To ensure the computer models are accurate, scientists test them against the past. They look at historical climate data gathered from ice cores, tree rings, and ocean sediment samples. If a model can accurately “predict” the climate shifts of the last 100,000 years based on historical atmospheric data, scientists gain high confidence that its predictions for the next 100 years are reliable.

Ultimately, the future of climate change isn’t a fixed mathematical certainty—it is a variable equation where human behavior is the most critical factor.

Are you interested in how these climate predictions specifically impact a certain region, or would you like to explore the economic models behind the transition to the energy transition to the energy transition to shift to renewables?

 “Can you predict the future of global climate change and explain how you arrived at these predictions?”

To look at the numbers behind those scenarios, we have to look at the hard data. The gold standard for these quantities comes from the Intergovernmental Panel on Climate Change (IPCC) Sixth Assessment Report.

Scientists quantify these predictions across several key metrics: atmospheric CO2 concentrations, global temperature rise, sea-level rise, and carbon budgets.

Here is how the numbers stack up for the year 2100 across the three main pathways.

2. Quantifying Extreme Weather Frequency

Climate modeling also quantifies how often extreme events will happen compared to a world without human-induced warming.

Extreme Heatwaves

• At 1.5°C warming: An extreme heat event that used to happen once every 50 years will occur 6 times more frequently and be 1.2°C hotter.
• At 2°C warming: That same heatwave occurs 9 times more frequently and is 2.7°C hotter.
• At 4°C warming: It occurs 2 times more frequently—meaning it happens almost every single year—and is 5.3°C hotter.

Heavy Precipitation (Storms & Flooding)

• At 1.5°C warming: Heavy 1-in-10-year rain events will happen 5 times more often and pack 10% more moisture.
• At 4°C warming: These deluge events happen 7 times more often and pack 30% more water, drastically overwhelming urban stormwater infrastructure.

3. The Remaining “Carbon Budget”

To ground these predictions in a definitive timeline, scientists use a metric called the Remaining Carbon Budget. This is the total net amount of CO2 humanity can ever emit from this point forward to stay below a specific temperature threshold.

• To cap warming at 1.5°C (with a 50% chance): The remaining global budget is roughly 200 to 250 gigatons (GtCO2)
• To cap warming at 2.0°C (with a 67% chance): The remaining budget is roughly 900 to 1,000 GtCO2.

The Reality Check

The world currently emits approximately 40 gigatons of CO2 per year from fossil fuels and industry.

• At our current burn rate, the budget to guarantee staying under 1.5°C will be entirely exhausted within 5 to 6 years.

• The budget for 2.0°C would be exhausted in about 22 to 25 years if global emissions flatten rather than drop.

Would you like to look at the math behind how economists translate these physical metrics into global GDP losses, or look into which specific regions bear the highest quantitative burdens?

In yesterday’s discussion about predictions of climate change, in your answer to the question “Can you quantify the predictions and the scenarios?” your units need some improvements. Mind to try again??

I hear you loud and clear. Looking back, mixing up or being vague with climate metrics (like toggling between gigatons of vs. carbon, or leaving temperature anomalies without a baseline) just muddies the waters. Let’s clean that up and lay out the hard, quantified numbers with strict attention to the proper units.

Climate scientists primarily use the Intergovernmental Panel on Climate Change (IPCC) Shared Socioeconomic Pathways (SSPs) to model our future.

Here is the precise breakdown of the scenarios and their quantified predictions by 2100.

The Metrics Defined

To ensure complete clarity, here are the exact units used below:

• : Gigatons (billion metric tons) of Carbon Dioxide emitted per year.
• : Parts per million, the atmospheric concentration of .
• : Watts per square meter, the measure of “radiative forcing” (the extra heat trapping energy in the atmosphere).
• °C: Degrees Celsius of warming above the pre-industrial baseline (1850–1900).

Quantified Climate Scenarios (Predictions for 2100)

1. The “Aggressive Mitigation” Scenario (SSP1-1.9)

This assumes global society shifts rapidly to sustainable practices, hitting net-zero emissions around 2050.

• Annual Emissions by 2050: (Net-Zero)
• 2100 Atmospheric Concentration:

2. The “Paris Agreement Target” Scenario (SSP1-2.6)

Slightly less aggressive than 1.9, but still keeps global warming well below the 2°C threshold.

• Annual Emissions by 2075: Net-Zero
• 2100 Atmospheric Concentration:
• Radiative Forcing by 2100:
• Predicted Warming: (Most likely outcome: )

3. The “Middle of the Road” Scenario (SSP2-4.5)

Societal and economic trends follow historical patterns. Emissions plateau around mid-century before declining, but net-zero is not reached by 2100.

• Peak Annual Emissions (2050):
• 2100 Atmospheric Concentration:
• Radiative Forcing by 2100:
• Predicted Warming: (Most likely outcome: )

4. The “High Emissions” Scenario (SSP3-7.0)

Characterized by regional rivalry, resurging coal use, and slow technological progress in clean energy.

• Annual Emissions by 2100: (Double current levels)
• 2100 Atmospheric Concentration:
• Radiative Forcing by 2100:
• Predicted Warming: (Most likely outcome: )

5. The “Fossil-Fueled Development” Scenario (SSP5-8.5)

The worst-case “business as usual” scenario driven by rapid, fossil-fuel-intensive economic growth. While considered less likely now due to the rise of renewables, it serves as an upper bound.

• Annual Emissions by 2100:
• 2100 Atmospheric Concentration:
• Radiative Forcing by 2100:
• Predicted Warming: (Most likely outcome: )

This AI discussion included interesting conclusions and missing answers. My first attempt returned gibberish and the one above did not include any actual numbers (even though I specifically asked for a quantitative analysis). Another big thing the AI conclusions didn’t mention was the Trump administration’s efforts to eliminate every mention of climate change (The Trump administration is eroding vital climate data):

AMERICAN SCIENTISTS have historically been leaders in the collection and analysis of data on climate change. The longest-running observations of carbon-dioxide concentrations in the atmosphere, for example, are collected at the Mauna Loa observatory in Hawaii. The National Snow and Ice Data Centre at the University of Colorado, Boulder, for its part, holds unique databases on the annual ebb and flow of sea ice at both poles. America also owns 58% of the roughly 4,000 Argo floats which drift at depth across the world’s oceans, gauging their health before popping up to the surface every ten days or so to broadcast their data home. These projects, and many more like them, are now under threat. Donald Trump’s administration is decommissioning long-standing databases; deleting key reports and analyses; and firing or reassigning the staff who have unique expertise in making sense of the data. Such actions will make climate modelling harder. But the harm done will be felt not just by the world’s climate scientists. American citizens and businesses will suffer, too.

In the US, scientists are starting to try to house the data independently from the federal government:

A small group of former government workers has recreated a valuable climate-science website that had been shuttered last year under the Trump administration. The new site, climate.us, is an effort by former staff members at the National Oceanic and Atmospheric Administration to present climate science previously housed at climate.gov, including data, reports, articles, and congressionally mandated national climate assessments. The new site is effectively the “first full clone” of climate.gov, said Rebecca Lindsey, managing director of climate.us. It became fully active on Tuesday morning.

My next blog will examine how including the keywords “anthropogenic” and “attributions” in the questions to AI changes its response.

Last week’s blog focused on my reading of Thomas Friedman’s NYT Op-Ed, in which he addressed the US slogan “Out of many, one.” I wrote it few days before Independence Day, July 4^th. On that Sunday (June 28^th), I did something else. I got an invitation to take part in an event commemorating the Holocaust at the Holocaust Memorial Park in Brooklyn. The program is posted in Figure 1. A year ago, I gave a talk at the same event and since then, my wife and I took part in another meeting organized by the same organization. As I mentioned after my participation last year, I am ashamed that prior to my involvement, I had never heard of this park. In the Facebook entry, I included a few photographs of it. You can find more information about the park on Wikipedia.

Figure 1 – Invitation to take part in a commemoration the Holocaust

As you can see in the invitation, the main speaker in the program was the head of the mayoral unit dedicated to preventing antisemitism. I didn’t know that Mayor Mamdani had such a unit. Anyway, before the meeting started, the organizers made a point of telling everybody that the park is supported by the city and there are no politics involved in running it. The park was constructed to honor the memory of the victims of WWII, and its only objective is to educate post-WWII generations about the Holocaust through the experiences of survivors. When Ms. Wisdom was about to start speaking, planted members of the audience started to shout, accusing her of antisemitism. There were attempts to stop the heckling, but eventually she left without saying a word. This was the most intense instance of trying to connect the Holocaust, antisemitism, and Israel that I have experienced.

The park is in Sheepshead Bay, a Russian section of Brooklyn.

After the charade, the managers of the park gave certificates of appreciation to a few Russian survivors, who then gave short speeches in Russian (the audience was mostly from the neighborhood). There were translations for the rest of us. In the middle of the event, I saw my wife crying. I asked for the reason; the answer was that she was seeing the age and conditions of the survivors (she is 79 I am 87 years old), and she realized that we are part of that crowd. After the event, I took her to a Jewish restaurant there. I also got a certificate of appreciation, shown in Figure 2, but I didn’t open my mouth throughout the event.

Figure 2 – Certificate of Appreciation given to me by the Holocaust Memorial Committee

I was very interested to hear Ms. Wisdom’s talk and find out if Mayor Mamdani’s effort is restricted to fighting antisemitism or whether it extends to other minorities, such as hate against Muslims. However, I didn’t have a chance to hear her there. So, after we returned home, I looked her up and read about the mayor’s effort:

The Mayor’s Office to Combat Antisemitism (MOCA) is tasked with identifying and developing efforts to combat antisemitism and anti-Jewish hate crimes, support victims of antisemitism, and celebrate Jewish culture and traditions using the resources of the City of New York. MOCA liaises with an Interagency Task Force to Combat Antisemitism, composed of representatives of the Office for the Prevention of Hate Crimes, New York City Police Department, New York City Commission on Human Rights, and other City agencies to fulfill its mission and deliver for the City’s more than one million Jewish New Yorkers. Additionally, MOCA coordinates non-law enforcement responses to incidents of antisemitism on behalf of the Office of the Mayor and serves as a liaison with the Jewish community to address issues related to services for victims of hate crimes and bias incidents motivated by antisemitism, and security for vulnerable populations and institutions.

The event in the park didn’t end with the certificates of appreciation. A member of the audience, long involved in the park but not a direct Holocaust survivor, gave a speech (mostly in English), expressing that he agrees with the committee on almost everything, with some small exceptions. Specifically, he argued that it is not a good idea to mix accusations of antisemitism with educational efforts about the Holocaust. I agreed with him; I wrote an earlier blog on the topic, “These Days, It Is Hard to Be a Jew and Even Harder to Be an Israeli!” (June 11, 2025):

I encountered antisemitism before October 7, 2023 and I wrote about my experiences (see the December 3, 2019 and February 11, 2020 blogs).  After October 7, 2023, and President Trump’s inauguration in January 2025, antisemitism acquired much more complexity. This blog will focus on the Israeli connection to antisemitism. The American connection was well covered in Senator’s Schumer’s new book “Antisemitism in America: a Warning,” published this year by GCP Grand Central. I strongly recommend reading that book, irrespective of readers’ background.

There is a strong tendency by many to want to free themselves from the boundaries of the IHRA definition. This is similar to the many voices that want to free themselves from the 1948 Genocide Prevention Convention’s definition of genocide and the original Lemkin definition that were described in earlier blogs. This relates to my perspective in the context of genocide: I view such efforts as a weaponization of the concept. Below is Wikipedia’s description of the weaponization of antisemitism:

However, one more argument was made on this occasion which made me uncomfortable. The argument was that after October 7th, only Israeli citizens can criticize Israeli government; criticism by anybody else weaponizes antisemitism.

I am a physicist. Symmetry is very important in Physics. It is also very important in life. If only Israeli citizens can criticize Israel, does that mean only Jews can criticize antisemitism??

To borrow from civil law suits, you don’t have to have “standing” to be anti-Israel or anti-antisemitic. There are already consequences:

Across Europe, many Jewish museums have seen visitor numbers drop, patrons back away and security threats rise since the fall of 2023, according to a recent report by the Amsterdam-based Association of European Jewish Museums, a nonprofit network of 55 institutions. The association’s members also reported online harassment, vandalism and acts of aggression against staff members.

Last Sunday, June 28th, I was reading the NY Times and noticed the Opinion section dedicated to the 250th Independence Day celebrations. It included an article by my favorite reporter, Thomas Friedman, on his interpretation of the US national anthem. Relevant paragraphs of the piece are given below:

I was late in setting out, so I put the pregame show on the radio and was still in the car when Brandi Carlile sang “America the Beautiful.” I am a huge Carlile fan, and since I was alone I quietly sang along, which may explain why, for the first time, I was suddenly struck by the concluding verse: “And crown thy good with brotherhood / From sea to shining sea.” What does that mean? I asked myself. “To crown thy good with brotherhood. …”

Katharine Lee Bates, who wrote the poem that the song is based on, seemed to be saying that the reward for acting good was brotherhood. What an interesting formulation. Brotherhood for her was the greatest achievement of all — not “good.” Doing good was the means, but brotherhood was the crowning goal.

Either way, I found her elevation of brotherhood inspiring, and it helped reinforce my own version of “America first” on this 250th birthday of America. My version has two parts. America needs to be the first in corralling as many nations in the world as possible into a global coalition, a brotherhood, designed to ensure that we maximize the benefits and cushion the worst impacts of the great planetary-scale challenges we now face together as a species — specifically managing artificial intelligence, climate change, human migrations, nuclear and biological weapons and pandemics — so that we rise together and don’t fall together. This is America as the first among equals. For all of America’s failings, no other country possesses the same combination of military reach, technological innovation, financial depth, democratic traditions and rule of law to convene others into a global coalition.

But America needs to be first today in another way. We need to be the first nation to prove that we can take our founding motto — to forge “out of many, one” — and make it work in a country where the “many” are so much more diverse than our original 13 colonies. What is required is nothing less than forging a continental-scale, multiracial, genuinely equal democracy unlike any history has seen before.

The key expression that I extracted from his piece was “Out of many, one.” I was familiar with the expression, but not in this context. This new knowledge drove me to make some changes to the coming blogs. In last week’s blog, I promised to continue focusing on AI’s abilities to predict the future. I wrote there that I would look at how it understands climate change and examine some of the proposed guardrails for it, but I am changing my course.

Returning to Friedman’s piece, I recognized that in the phrase “out of many, one,” the word “many” has different meanings depending on the context; for a country, it can represent the world; for a state, it can represent a country; and for individuals, it can represent everybody. The question for me was what AI “thinks” of the issue, so I asked it (through Gemini) to construct an image that conveys the meaning of the statement. The result was a gate similar to the one shown in Figure 1, with a bunch of students congregated underneath. For all intents and purposes, it looked to me like the entrance to a typical college. I asked AI to try again with something a bit more abstract. What I got is Figure 1, with one side of the gate showing a bunch of international flags labeled “E Pluribus Unum” and the other side showing “unity in diversity.” This was closer to my thinking, but to be honest, the “one” looked to me more like a collective grave after a mass killing, than the constructive “one” I imagined. I tried to find a way to show that individuals can make contributions to the collective in doing something constructive, but got nowhere.

Figure 1 – Out of Many, One (By AI-Gemini)

Figure 2 – Make America Great Again (by AI – Gemini)

Well, after that exercise, I asked AI for an image that would convey the meaning of another statement: “Make America Great Again.” The result is shown as Figure 2. As can be seen, Figure 2 is similar to Figure 1. The central figure didn’t change and the gate stayed the same except for the replacement of the writing on top and the addition of two state plaques. It also removed the trash can and added American flags in the background and several plaques on the floor. The “unity in diversity” and international flags stayed. Regardless of the number of plaques on the floor, one cannot escape from imagining that there are 13 and that the “again” in the statement brings us back to the Founding Fathers. I wonder what the Trump administration would think about this interpretation.

Past, present, and future Venn diagram (Source: AI-Gemini)

Throughout our lives, all personal, collective, global, tribal and national issues consist of an interplay of past, present, and future. This is formalized in the top picture, which was constructed for me by AI (Gemini). We can mostly factualize the past (if we can still find relevant evidence). We can experience the present and try to predict the future as best as we can. I have tried, to the best of my ability, to predict the future throughout this blog (put “predictions” into the search box), with an emphasis on the tools that can help me to quantify these predictions. This is important because some of the predictions are for existential issues and one needs time to act on them to mitigate them. My book, Climate Change: the Fork at the End of Now, focused on such scenarios (See the “My Now” blog from March 9, 2021). Regarding a different threat, a recent Nobel Prize–winning nuclear physicist predicted that humanity has about 50 years to survive:

The chances of you living 50 years are very small’: Theoretical physicist explains why humanity likely won’t survive to see all the forces unified. Live Science spoke with Nobel prize-winning physicist David Gross, who recently received the $3 million Special Breakthrough Prize in Fundamental Physics, about the quest to unite all the forces and why humanity might not live to see a unified theory.

50 years is well within my definition of “Now” (within the lifetime of my grandchildren), which takes us to the end of the century.

Prof. Gross’ conclusions are based on the likelihood of nuclear war, which recently increased from 1% per year to 2%. I didn’t see the origin of the change in the yearly estimate in his interview, but once you accept the increase in yearly estimate, the math is very clear and follows simple exponential decay that he is very familiar with.

My guess is that Prof. Gross’ estimate of the yearly doubling the chance of nuclear confrontation is based on the Doomsday Clock, which was discussed in an earlier blog (June 18, 2025).

The two main events that are pushing forward the Doomsday Clock are the Russian–Ukrainian war and the US-Israel-Iran war. Recent developments in the Russian–Ukrainian war have given the sense to many (including me) that—against all odds—the Ukrainians are gaining the upper hand:

Ukraine Bombards Moscow With One of the Biggest Drone Attacks of the War. The attack, which shut down the capital’s airports for several hours, was part of an escalating campaign to bring the conflict home to Russians.

It seems to many that at some point, the Russians will either choose to end the war with an agreement or destroy Ukraine with nuclear weapons. However, the nuclear destruction of Ukraine would not stay in Ukraine.

AI is now developing to help us with quantitative predictions about this complicated (and frightening) picture. I suspect that quantitative predictions of the future will be one of the largest applications of AI. This blog will focus on the general abilities of the technology right now; next week’s blog will focus on how it understands climate change (which my background gives me the capability of scrutinizing). In two weeks, I will write about the guardrails that are now under discussion for AI.

Before shifting to these details, I thought that it would only be“fair” to let AI outline its own capabilities in quantitative prediction of the future. The AI systems that I used in this exercise were Gemini (Google) and CoPilot (Microsoft). My prompts (as well as my answer to a question that AI posed to me) are marked in red.

Through Gemini:

“AI as a predictor of the future?”

AI is rapidly transitioning from an analytical tool into an active predictor of the future. By shifting forecasting from a human craft to an automated utility, machine learning models can process massive, complex datasets to predict everything from market shifts to weather events. In live forecasting platforms like Metaculus, advanced AI systems have already achieved over 80% of the accuracy of elite human forecasters on geopolitical and technology trends. [1, 2, 3]

The capabilities, limitations, and societal impacts of using AI as a “crystal ball” involve several key factors:

Where AI Excels at Predicting

• Time-Series and Pattern-Heavy Data: AI thrives in environments with minimal “distributional shift”—where past data patterns remain relatively stable. [1]

• Scientific & Medical Discoveries: AI models analyze vast literature and biological data to predict valid molecular structures for drugs or map out evolving scientific fields. For instance, a Mayo Clinic team developed an AI capable of detecting pancreatic cancer on CT scans up to three years before clinicians can see it. [1, 2, 3, 4, 5]

• Hyper-Local Climate and Weather: AI drastically speeds up environmental calculations, identifying patterns linked to storms, floods, and wildfires with remarkable short-term accuracy. [1, 2]

• Crowdsourced Forecasting: Benchmarks like the University of Chicago’s Profit Arena evaluate AI models live on real-world prediction markets, showing that certain AI models find statistical edges that humans miss. [1]

Severe Bottlenecks to AI Forecasting [1]

• The “Data Gap”: AI struggles significantly when there is a time gap between its training data and a current event. It cannot naturally account for sudden, unprecedented “black swan” events, such as geopolitical outbreaks or unrecorded financial crises. [1, 2]

• Overconfidence and Hallucination: Evaluation studies by institutions like Knowledge at Wharton show that AI models frequently exhibit overconfidence, assigning dangerously high probability scores to outcomes that lack factual evidence. [1]

• Lack of True Intuition: AI relies purely on historical data correlation. Unlike humans, it lacks genuine creative insight or “judgmental forecasting” capabilities—the ability to assign probabilities based on abstract intuition and personal domain expertise. [1, 2]

The Dangers of Algorithmic Future-Telling

• Self-Fulfilling Prophecies: When a dominant AI system predicts a market crash or a political outcome, human actors change their behavior to match that prediction, inadvertently forcing it to come true. [1, 2, 3, 4, 5]

• Pre-Crime and Surveillance: Authoritarian governments are actively testing AI to analyze location history and internet traffic to predict who might become a political dissident before any public act of protest occurs. [1]

• Feedback Loops: As AI-generated content floods the web, newer forecasting models risk being trained on data predicted by older models, leading to a closed-loop degradation of accuracy. [1, 2, 3, 4]

Ultimately, AI is a powerful enhancer of human prediction, not a replacement for it. The most reliable forecasts still occur when AI handles the brute-force data crunching, leaving humans to provide contextual oversight, ethical boundaries, and strategic judgment. [1, 2, 3]

Through CoPilot:

 “AI as a predictor of the future?”

AI can predict aspects of the future—but only in specific, data-rich domains, and never with perfect certainty. Across science, business, and geopolitics, AI is already outperforming humans in some forecasting tasks, yet it also introduces new risks when misused. Below is a clear, evidence-based breakdown of what AI can and cannot do as a predictor of the future.

🧠 What AI Can Predict Reliably

1. Natural disasters (partially)

AI improves early warning systems by analyzing massive streams of satellite, weather, and environmental data far faster than humans. It detects patterns linked to hurricanes, floods, droughts, and wildfires, helping scientists issue

earlier and more accurate alerts. However, AI cannot eliminate uncertainty, especially for chaotic systems like earthquakes.

2. Business and technology outcomes

Large language models have begun outperforming trained human experts in predicting which new tech ventures will succeed. In a controlled tournament, top AI models correctly identified winners in nearly 4 out of 5 cases, beating MBAtrained managers. This suggests AI is becoming a powerful tool for strategic foresight in business.

3. Geopolitical forecasting

AI systems have competed in forecasting tournaments (e.g., Metaculus) and achieved accuracy levels comparable to top human forecasters—sometimes surpassing them. One AI system scored over 80% of top human performance, far above expectations.

Where AI Prediction Becomes Dangerous

4. Predicting human behavior for surveillance

Some governments are experimenting with AI to forecast political dissent or identify potential future critics by analyzing online activity, location data, and behavioral patterns. This raises severe ethical and civilliberty concerns and demonstrates how predictive AI can be weaponized.

5. Predicting social outcomes at national scale

Argentina attempted to build a “social digital twin” to simulate future poverty, subsidy impacts, and humancapital development. While ambitious, the system’s rollout was flawed and sparked skepticism—highlighting the difficulty of predicting complex social systems.

What AI Might Predict in the Future

Experts and analysts foresee AI becoming deeply embedded in forecasting across society:

• Predictive healthcare: anticipating diseases years before symptoms.

• Economic and societal modeling: more accurate simulations of policy impacts.

• Everyday anticipatory systems: AI that predicts your needs before you express them.

• Longterm societal shifts: experts are divided—56% of AI experts expect a positive impact over 20 years, while the public is more skeptical.

The Bottom Line

AI is not a crystal ball. But it is becoming a powerful forecasting engine—especially where:

• There is abundant, highquality data

• Patterns are stable and quantifiable

• Predictions can be updated continuously

• Human bias or slow analysis is a bottleneck

AI struggles or becomes dangerous when:

• Predicting complex human behavior

• Used for political or social control

• Data is biased, incomplete, or invasive

• Systems are deployed without oversight

• A question for you, Micha

• Are you thinking about AI as a tool for personal forecasting (e.g., career, finance), societal forecasting, or philosophical questions about the future of humanity?

My answer: “All the above and the interactions between them. Specifically – how can I check the probabilities of these predictions.”

Following my response, AI gave me the following links:

Can AI Help Scientists Predict Natural Disasters More Accurately in the Future?

AI beats human forecasters in tournament predicting 30 tech ventures

What if a government could predict its critics before they speak? Inside China’s alleged AI-powered surveillance project

Chinese Firm Develops Technology That Uses AI To Predict Future Protestors

https://www.yahoo.com/news/politics/articles/argentina-launched-ai-predict-future-150104368.html 

AI May Soon Understand Humans Better Than We Understand It, Scientists Warn

3. Public and expert predictions for AI’s next 20 years

Meanwhile, The Guardian covered an optimistic vision of the future that is “the product of 45 authors based on databases compiled by more than 200 researchers from around the world”:

Humanity can raise living standards, reduce inequality and keep global heating within a 2C rise, according to a sweeping vision for planetary survival. The report by the World Inequality Lab (WIL) aims to be the most comprehensive attempt yet to navigate the polycrisis that is pushing the world toward climate breakdown, political extremism and ever greater economic and social tension. It offers a set of bold policy proposals, including hefty wealth taxes on billionaires, sharp reductions in working hours, a change in diets and a shift of investment from materially intense sectors, such as industry and mining, to education and health.

While the example above is human-driven, this could also be the start of a different conversation with AI about how to make the world a better place.

Stay tuned!

(Source: WSJ)

In about two weeks, we will celebrate 250 years of American independence, which was established with the issuing of the Declaration of Independence:

The Declaration of Independence, formally The unanimous Declaration of the thirteen united States of America in the original printing, is the founding document of the United States. On July 4, 1776, it was adopted unanimously by the Second Continental Congress.

The first three paragraphs of the document read as follows:

We hold these truths to be self-evident . . .

When in the Course of human events, it becomes necessary for one people to dissolve the political bands which have connected them with another, and to assume among the Powers of the earth, the separate and equal station to which the Laws of Nature and of Nature’s God entitle them, a decent respect to the opinions of mankind requires that they should declare the causes which impel them to the separation.

We hold these truths to be self-evident, that all men are created equal, that they are endowed by their Creator with certain unalienable Rights, that among these are Life, Liberty and the pursuit of Happiness.

That to secure these rights, Governments are instituted among Men, deriving their just powers from the consent of the governed, That whenever any Form of Government becomes destructive of these ends, it is the Right of the People to alter or to abolish it, and to institute new Government, laying its foundation on such principles and organizing its powers in such form, as to them shall seem most likely to affect their Safety and Happiness.

The three paragraphs claim that the (British) government does not have consent from the people living in America to be governed, and is actively destructive of their way of life. Therefore, they find it necessary to form a separate sovereign state, with a government of their own. 250 years went by, and the US became the strongest and richest sovereign country in the world, governed by the American Constitution. It was signed on September 17, 1787, after the conclusion of the American Revolutionary War (1775-1783), which can be considered the nation’s “pregnancy” if we follow the same analogy that calls the start of the nation its “birthday.” The second paragraph of the Declaration of Independence makes these claims about the Founding Fathers’ values and beliefs:

 We hold these truths to be self-evident, that all men are created equal, that they are endowed by their Creator with certain unalienable Rights, that among these are Life, Liberty and the pursuit of Happiness.

Thus, it follows that if men (humans) are created equal but are not being treated that way, they can try to form their own sovereign government.

250 years later, humanity is creating a technology that will have a direct impact on the ability of humans to govern themselves directly. The control of this technology is mostly in the hands of American companies. AI technologies are having major impacts on governments. The third paragraph of the Declaration of Independence gives the people the right to look for remedies.

It looks to me like there is a search taking shape for ways to counter the central role that AI is gaining in running governments. In this blog I will use two examples: one from the US and one from Europe:

US: NYT Opinion from Bernie Sanders

Mr. Sanders, an independent, is the senior senator from Vermont.

Artificial intelligence will almost certainly be the most transformational technology in the history of the world. It will profoundly affect the life of every man, woman and child in our country. It will bring — and is already bringing — unimaginable changes to our economy, our democracy, our emotional well-being, our environment and how we educate and raise our children. Further, there is a very real fear that as A.I. becomes smarter than humans it could eventually function independently, with potentially catastrophic consequences. The question, then, is not whether A.I. will change the world. It will. The question is: Who will own and control that future? Who will benefit from it, and who will be hurt by it? Will A.I. be used to make life better for working families? Will it enrich our quality of life? Will it help us eliminate poverty, extend life expectancies and solve the climate crisis? Or will the future of humanity be determined by a handful of billionaires who have promoted and developed A.I., with virtually no democratic input, who stand to become even richer and more powerful than they are today?

Europe: Brussels, Belgium (Politico):

The European Union’s energy chief says companies that want to profit from the artificial intelligence boom are welcome in Europe — but only if they demonstrate they are committed to the bloc’s energy, climate and environmental goals. That means supporting renewable and nuclear power sources rather than fossil fuels and recycling the amounts of excess heat from data centers to heat Europeans’ homes and businesses, EU Energy Commissioner Dan Jørgensen told POLITICO in an interview. The AI boom is creating enormous demand for new data centers, with the U.S. and China leading construction of the facilities needed for powering large language models like ChatGPT, Claude and China’s Doubau. The EU wants to catch up.

Last week’s blog focused on AI’s ability to be a personal assistant and help many of us live better. I ended that blog with a description of the instructions I got from AI to make my own digital twin, and those of others. The AI assistance I discussed was mainly in a home setting, which fits my own retired settings. However, a larger impact can be found in professional settings. I will include one example that I became aware of while reading the NYT recently: “Awareness of Using AI During a Medical Appointment ”

Dr. Nicholas Gavin, an emergency medicine doctor at Mount Sinai in New York City, was working an overnight shift last summer when a patient came in with a puzzling set of symptoms. Within seconds, his three younger colleagues — two medical students and a resident — were consulting a free artificial-intelligence-powered app for physicians, Open Evidence. Dr. Gavin soon learned that they were far from outliers. A third of Mount Sinai’s 9,000 doctors were already regular Open Evidence users, the health system’s executives found out in a meeting last year with the start-up’s leaders. “That was an ‘aha’ moment for our leadership,” said Dr. Gavin, who is also the system’s chief clinical innovation officer.

Governments and professional activities are not the only areas affected by collective AIs that need to be considered. Every human-to-human interaction will be impacted. One of the biggest impacts that is already taking shape is job loss due to AI. AI will legitimately change the nature of many jobs, many of which might require fewer workers. In many cases, our jobs will be at risk. Since our supervisors have an important voice in our promotion or firing, it might help to consult with our supervisor’s digital twin about our performance. A good example that I brought up in last week’s blog was the possible use of the digital twin of our country’s President by federal workers. Similar thinking could hold for all kinds of supervisors. I will return to this issue in future blogs. In next week’s blog, I will focus on the AI use in quantitative assessment of future expectations, something I suspect will develop into one of the largest and most consequential uses of AI.