The climate case for AI has not been made

This week, as countries meet in Antalya, Turkey for the United Nations climate conference, one topic likely to dominate the conversation is artificial intelligence. The host has already called on countries to commit to ensuring that AI is designed, powered, and deployed in ways that support climate action. But so far, industries around the world aren’t doing nearly enough to harness the technology to actually drive down the greenhouse-gas emissions heating up the planet, despite some very optimistic claims from AI companies. In 2024, OpenAI CEO Sam Altman declared that AI would unleash “astounding triumphs,” including “fixing the climate.” The year before, in a 2023 op-ed, the head of sustainability at Google and her coauthor wrote that AI could reduce global emissions 5% to 10% by 2030. Some academic researchers have examined these AI-related climate claims. A peer-reviewed study published last year, led by researchers at the London School of Economics’ Grantham Research Institute and the firm Systemiq, estimated that AI could help cut global emissions by the equivalent of 3.2 to 5.4 billion tons of carbon dioxide annually by 2035—roughly on par with the US’s annual energy-related emissions on the high end, even after subtracting AI’s own energy footprint. Those potential gains are mostly concentrated in three sectors: power, meat and dairy, and vehicles. But projections like these are only as good as our willingness to test them against present-day reality. So is there evidence that the largest companies in these three sectors are beginning to integrate AI in line with the study’s projections, and if so, do the results match expectations? Or is the 2035 number still mostly aspirational? Companies across these sectors have announced plenty of initial efforts to incorporate AI into various business practices. But an AI-assisted review of public information from significant market actors in each sector, conducted and reviewed by the second author of this piece, found that few have announced they’re pursuing the specific emissions-reducing routes that informed the projections in the study—and fewer still claim they’ve used them to achieve meaningful climate progress. The power sector is where the earlier study’s underlying math is most concrete: AI-driven forecasting of supply and demand could increase the amount of renewable electricity that reaches the grid from already-built solar and wind facilities by up to 20%, which could prevent an estimated 1.8 billion tons of emissions a year. Yet of the grid operators reviewed in China, the United States, and Europe, only one deployed a known meaningful, working AI program of this sort, according to the analysis. A partnership between the UK’s National Grid ESO and Open Climate Fix halved the grid operator’s solar-forecast error (the difference between the amount of incoming sunlight predicted and the amount that actually reaches solar panels), avoiding hundreds of thousands of tons of carbon dioxide a year. Beyond this example in the UK, though, many other grid operators in the EU, the US, and China had little to show in the public record in terms of the AI integration envisioned by the study. The meat and dairy sector offers the starkest contrast between technology readiness and market reality. To achieve the 2035 emission reductions anticipated in the earlier study, AI would need to close the taste and cost gap between traditional proteins and alternative ones, such as lab-grown or plant-based burgers and milk substitutes. In one of the clearest efforts to achieve this, the AI and food tech company NotCo recently formed a joint venture with Kraft Heinz using its platform Giuseppe, which analyzes plant molecular data to pinpoint combinations replicating the taste, texture, and smell of animal products. But few other alternative-protein companies had such a deeply integrated way of using AI in product development, the AI-assisted review suggests. Meanwhile, the sector seems to be facing more fundamental problems, including stubbornly slow mainstream consumer adoption and declining funding levels. Most notably, venture capital flowing into cultivated meat fell from a $989 million peak in 2021 to around $83 million last year. The transportation sector offers some hopeful examples of the type of AI integration necessary to achieve the study’s projections, particularly among battery manufacturers and ride-sharing platforms. CATL—which controls roughly 40% of the global market for electric-vehicle batteries—launched an AI system trained on more than 100,000 battery design cases and 600 terabytes of test data. It cut the time needed to screen candidate materials by two orders of magnitude, according to the company. Independent researchers also found that the ride-sharing platform Didi Chuxing was able cut per-trip emissions 16% on average by using AI-optimized routing to minimize detours. There’s another big problem in the climate case for AI, though: Oil and gas companies are also interested in these tools for many reasons, including the potential to increase production of fossil fuels. Technology firms that committed to sustainability goals in the past should answer questions about how their tools could be used to boost old, dirty energy as well as clean, renewable energy. Finally, it’s also worth noting that integrating AI is just the first step, and it remains unclear to what degree the technology can accelerate the sorts of improvements and innovations envisioned. AI also does little to shorten the standard paths to certifying, permitting, building, packaging, marketing, and selling any resulting products, facilities, or services. The findings of the analysis, carried out by Anthropic’s Claude Code, were reviewed against public documents and media accounts to ensure their accuracy. It is possible that some companies and utilities in these industries are pursuing AI-assisted applications that would fit into the initial study’s scenarios but haven’t yet been disclosed publicly. Greater and better-targeted AI integration and efficiency gains in these three sectors may well materialize, bringing the 2035 projections closer to reality. But it won’t happen without deliberate and sustained efforts to ensure that AI delivers more environmental benefits than harms. Notably, so far the rapid adoption of AI models and buildout of data centers has driven a huge increase in energy consumption and emissions, and this has led major tech companies like Microsoft, Google, and others to fall well off track in their corporate climate goals. At this point it would be surprising if the sector even manages to zero out its climate pollution: Emissions associated with global data centers (which may handle AI as well as non-AI computing tasks) could exceed 300 million tons annually in 2035, according to the IEA’s base forecast. Industry leaders must implement the technology in the ways that promise to do the most to drive down emissions. Corporate customers should select models with the lowest environmental impacts and opt to work with the AI developers doing the most to reduce their climate footprint. And states and regions need to do more to incentivize or require all these companies to operate in more responsible ways.