news

last week, openai ceo sam altman said in an article that the rapid improvement of ai capabilities will usher in an ideal "intelligent era" that will bring "unimaginable" prosperity and "amazing achievements" such as "solving climate change". question".

but this is a vision that no one can commit to, especially when it comes to climate change, and it fundamentally misunderstands the nature of the problem.

even more irritating, this argument implies that the massive amounts of electricity currently consumed by ai are irrelevant because it will help us generate abundant clean energy in the future. that argument downplays growing concerns about the technology, which is already driving proposals to build natural gas power plants and diverting big tech companies from their climate goals.

there is no doubt that the energy demands of ai will only continue to increase, even as the world scrambles to build bigger, cleaner power systems to meet the needs of electric vehicle charging, green hydrogen production, heat pumps and other low-carbon technologies. altman himself is said to have just recently met with white house officials to lay out plans to build absolutely massive ai data centers that would likely require the equivalent of five dedicated nuclear reactors to run.

the basic point that mit technology review has always adhered to is that technological progress can indeed bring practical benefits and accelerate social progress in meaningful ways. but for decades, researchers and companies have overhyped the potential of ai, claiming it could lead to breakthrough medicines, enable superintelligence, and even free humans from work demands. to be fair, there have been some notable developments, but nowhere near the level of hype.

given this record, i think before you can claim that ai can solve humanity's toughest problems, like widespread poverty or global warming, you need to develop a tool that does more than plagiarize news reports or help students cheat.

of course, ai may help combat the growingclimate changethreaten. we’re already seeing research teams and startups using this technology to manage the grid more efficiently, quickly fight wildfires, and discover materials that can make cheaper, better batteries or solar panels.

these advances remain relatively incremental. but even assuming that ai does deliver an energy miracle, perhaps its pattern-recognition capabilities will provide key insights that could ultimately crack nuclear fusion—a technology that altman, as an investor, is betting heavily on.

this will be a very good thing. but technological progress is just the beginning—eliminating global climate emissions is far from achievable by technological progress alone.

why do i say that?

because from nuclear fission plants, to solar farms, to wind turbines to batteries, we already have all the technology we need to clean up the power sector. this should be the easiest part of the energy transition. yet in the largest economies on earth, fossil fuels still generate 60% of the electricity. that so much of our electricity still comes from coal, oil and gas is not just a technical problem but a regulatory failure.

“as long as we continue to effectively subsidize fossil fuels by allowing them to use the atmosphere as a dumping ground for waste, we cannot allow clean energy to compete on a level playing field,” zeke hausfather, a climate scientist at the independent research organization berkeley earth, responded to altman on the x platform ’s article was written. “we need policy changes, not just technological breakthroughs, to achieve our climate goals.”

that’s not to say we don’t need to solve significant technical challenges. just look at our ongoing efforts to develop clean and cost-competitive crop fertilizers or aircraft fuels. but the fundamental challenges of climate change are sunk costs, barriers to development and inertia.

we have built and paid for a global economy that emits greenhouse gases, investing trillions of dollars in power plants, steel mills, factories, jets, boilers, water heaters, stoves and suvs that run on fossil fuels. as long as these products and factories are still running, few people or companies are willing to give up these investments easily. ai won’t solve all of this simply by generating better ideas.

to dismantle and replace machinery in industries around the world at the pace currently required, we need increasingly stringent climate policies that incentivize or force everyone to move to cleaner factories, products and practices.

but pushback arises every time tougher laws or a big new wind or solar project are proposed because the plans would affect someone's wallet, block someone's view, or threaten someone's cherished values. region or tradition. climate change is an infrastructure problem, and building infrastructure is a messy human activity.

technological advances can alleviate some of these problems. cheaper, better alternatives make hard choices more politically attractive. but no improvement in ai algorithms or underlying data sets can solve nimbyism, human conflicts of interest, or the desire to breathe fresh air in unspoiled wilderness.

claiming that a single technology—one that happens to be developed by your company—can miraculously unravel the conflicts entrenched in these societies is self-serving at best and naive at the very least. and it’s a worrying argument at a time when the growth of this technology is threatening the little progress the world has made on climate change.

for now, we can safely say that generative ai is making the most difficult problems we face even more difficult to solve.

original link:

https://www.technologyreview.com/2024/09/28/1104588/sorry-ai-wont-fix-climate-change/