How China became climate-pilled--and what that means for AI

AI safety needs to be seen as a domestic priority for Chinese policymakers, not a foreign demand. Climate change offers a useful parallel.

Clean tech superpower

China is a clean-tech superpower today, manufacturing and deploying most of the world’s clean-tech goods from electric vehicles and batteries to solar panels and wind turbines. I have often argued that China went all in on clean tech because it had to. As the world’s largest importer of oil and gas, China has long been worried about energy security and sees clean energy as the ultimate solution.

But China’s clean tech obsession is not just about energy security. China is genuinely climate-pilled. Chinese policymakers genuinely believe that climate change is real, that the effects are serious, and that as the world’s largest carbon emitter China can and should do something to stop it.

If China were merely concerned with its own energy security and not climate change per se, then it wouldn’t be making such an enormous effort to shift away from coal, which it has in abundance. Nor would it be working so hard to reduce emissions in industrial sectors such as steel, aluminum, and cement.

The world is lucky that China is now climate-pilled. But this wasn’t always the case. I argue that China reached a turning point in the late 2000s when Chinese leaders really internalized the climate issue rather than seeing it as mostly someone else’s problem.

Chart: Kyle Chan. Data: Ember.

Taking AI risks seriously

Today, there’s a new major global risk: AI. While AI is a powerful tool that can be used to improve people’s lives, it also creates new significant risks from AI-enhanced cyberattacks to biological risks to runaway AI systems, like the recent OpenAI incident with Hugging Face. Up until recently, Chinese policymakers have not taken these large-scale risks as seriously, instead focusing more on political and social risks such as deepfakes and AI labelling.

Some of this is due to the fact that Chinese policymakers are not AGI-pilled, as I’ve written about, and do not see revolutionary superintelligent AI systems as being just around the corner. But China doesn’t need to be AGI-pilled to take large-scale AI risks more seriously.

As we think about global AI governance and AI safety today, it’s helpful to look back at China’s shifting attitude toward climate for lessons on how to get Chinese policymakers to take major AI risks more seriously—on their own terms.

Climate as a trap

For a long time, China pushed back on outside efforts to get it to act faster on climate change. China’s main concern back then was getting enough energy to power its rapidly growing economy. Blistering growth in manufacturing in the 1980s and 1990s constantly threatened to overwhelm China’s fragile and outdated power grid. Peak power demand frequently exceeded supply by 10 to 20 percent, and rolling blackouts were common in parts of the country.

Source: Our World in Data

Officially, China took climate change seriously from the beginning but was worried about binding commitments that would slow down its own development. China was a founding member of the Intergovernmental Panel on Climate Change (IPCC) when it was established in 1988. But in many rounds of climate negotiations, China pushed back against hard commitments, arguing that industrialized nations were primarily responsible for historical emissions and should thus bear more of the costs of mitigation.

In a 2007 speech, President Hu Jintao famously argued:

无论从历史责任还是从现实能力而言，发达国家均应率先减排，并在减缓和适应气候变化方面向发展中国家提供帮助。对于广大发展中国家而言，发展经济和改善民生是第一要务。

Whether in terms of historical responsibility or present-day capabilities, developed countries should take the lead in reducing emissions and provide assistance to developing countries in mitigating and adapting to climate change. For the vast majority of developing countries, the foremost priority is economic development and improving people’s livelihoods.

Data: Our World in Data

Unofficially, Chinese commentators were even pushing the idea of a “low-carbon conspiracy” (低碳阴谋) where the West was using climate change to deliberately slow down China’s development. While far from mainstream, there was a prominent strain of climate denialism in China with some influential voices calling it a hoax (see this great 2017 Foreign Policy piece by Geoff Dembicki).

Interestingly, this stopped after 2011 when China made clean tech a core part of the 12^th Five-Year Plan, and questions remain about whether China might have censored climate denialism outright.

Climate’s threat to China

A major turning point in China’s attitudes toward climate change occurred in the mid-2000s. Beijing launched a major government study looking into the impact of climate change on China that spanned twelve government agencies over four years. In 2006, this study was published as the “National Assessment Report on Climate Change” 《气候变化国家评估报告》and found that climate change was already threatening China’s fragile ecosystem and food security through worsening droughts, floods, and extreme weather.

In 2007, China launched its first National Climate Change Program 《中国应对气候变化国家方案》which laid out a multi-pronged strategy for reducing emissions, shifting to renewable energy, and preparing adaptation measures for agriculture and other sectors. Importantly, this plan was created by China’s top economic agency, the National Development and Reform Commission, which meant this was being incorporated directly into China’s core development efforts rather than being treated merely as a side matter.

During this time period, climate went from being a foreign relations issue for China to becoming a core national security concern. In particular, the threat to food security, long one of China’s greatest anxieties given its limited arable land and history of severe famines, made climate a serious and urgent domestic challenge requiring the attention of top political leaders.

When Xi Jinping made his big “dual carbon” pledge in 2020 to peak carbon emissions by 2030 and become carbon neutral by 2060, this was not something China was pressured or forced into but something that China ultimately saw as being in its own strategic interests.

Source: The Economist

Air pollution and “airpocalypse”

Another major factor driving China’s changing stance on climate was the air pollution crisis. For years, Chinese scientists had warned of the public health risks from hazardous levels of air pollution caused by industrial activity, coal power plants, and other sources. Chinese cities such as Beijing had been monitoring and trying to cut down on particulate matter, sulfur dioxide, and nitrogen oxides since at least the 1990s.

But the air pollution issue was kicked into high gear when the US Embassy in Beijing began publishing air pollution data on Twitter in 2008. It revealed to the public that the pollution problem was far worse than the official Chinese data made it seem.

At first, Beijing pushed back, calling the US Embassy’s pollution tweets unscientific and illegal. But over time, China actually started to change and established a nationwide network of 500 air monitoring stations in over 70 cities.

This culminated in 2014 when Li Keqiang declared “war on pollution” (宣战) following a particularly severe stretch of smog in 2013 that was dubbed the “airpocalypse.” Air pollution and other environmental measures were incorporated into promotion criteria for local government officials, and Chinese cities have made significant progress in combating air pollution, although more work is clearly needed.

Chart: Kyle Chan. Data: US Embassy in Beijing. See my Github

While air pollution is a separate issue from climate change, the two are linked through coal, industrial emissions, and vehicle exhaust. Perhaps most importantly, the air pollution issue helped to bring home the impact of environmental harm and get China to internalize climate and environment as critical domestic issues rather than as something pushed onto them by other countries.

Now, so-called “ecological civilization” (生态文明) has become central to the Chinese government’s own conception of what it means to be a modern and developed country.

Clean tech growth engine

Around the same time, China began to shift from seeing climate as a constraint on development to a new engine for economic growth, i.e., the so-called “new energy industries” (新能源产业).

This shift was partly driven by China’s surprising boom in solar manufacturing during the early 2000s in response to policy incentives from other countries, like Germany (see my piece on Chinese solar and the German paradox). It was also driven by visionary policymakers like Wan Gang, a former Audi engineer who became Minister for Science and Technology and the godfather of China’s EV program.

Chart: Kyle Chan. Data: Earth Policy Institute, IEA-PVPS.

In 2010, China identified “new energy industries” such as solar, wind, and EVs as part of a set of “strategic emerging industries” that it would target for industrial policy. Clean tech became central to China’s broader efforts to upgrade existing industries and “seize the commanding heights of a new round of economic and technological development” (抢占新一轮经济和科技发展制高点的重大战略).

Clean tech and the transition to renewable energy became enshrined in the 12^th Five-Year Plan and has been a key part of China’s Five-Year Plans ever since (see my piece on China’s tech long game). Today, EVs, solar, and batteries are touted as China’s “new three” (新三样) export growth drivers.

Chart: Kyle Chan. Data: Ember.

The climate and AI parallel

There’s an interesting parallel between China’s views on climate and large-scale AI risks. With climate change, China underestimated the scale of the risks, and some even believed it was a plot to undermine China’s development. But over time, China came to internalize the risks of climate change and is now charging ahead with the energy transition. In the meantime, the tables have flipped with Trump famously tweeting in 2012 that climate change was a hoax created by China to undermine American manufacturing.

A similar story could unfold with AI safety. For now, Chinese policymakers are more focused on near-term political and social risks from AI, such as deepfakes and child safety. But they haven’t taken large-scale AI risks, such as rogue agents and loss of control, as seriously. But this is starting to change.

As outsiders, the US can nudge and encourage China’s shifting stance toward AI safety. But we shouldn’t use coercion or try to strike a grand bargain. These will backfire and reinforce a view among some in China that the US is using AI safety as an excuse to slow China down—like the way climate was once framed as a Western plot to hold back China’s development.

Ultimately, China needs to take major AI risk seriously on its own, just like it did with climate change. Only then will China tackle the problem in a robust rather than a superficial manner. ■