Here is how to deal with it
Illustration: Paul Blow
Advances in ARTIFICIAL intelligence have long terrified techies. Lately, voters are feeling the angst, too. AI is unpopular in the West and climbing up the political agenda. The fiercest fights so far have been in America, where protests against data centres have scuppered nearly $100bn-worth of projects, warring AI megadonors have just dumped tens of millions into a Manhattan congressional race and around 40% of voters tell pollsters that they want AI banned from most industries. But spats are breaking out elsewhere: after chipmaking profits soared recently, workers at Samsung in South Korea threatened a strike to secure special payouts.
The backlash is only just getting started, because the technology is only just getting started, too. Britain’s flimsy prime-minister-in-waiting, Andy Burnham, has barely said a word about AI. Even Americans still rank it 29th out of 39 election issues.
That is bound to change—and battles over data centres offer a hint of the struggles to come. The buildings summon a vitriol well beyond conventional nimby ism. More Americans say they would be happy with a nuclear reactor next door than a data centre. Even plans to build one in the Utah desert have met with passionate opposition.
Data centres can be ugly, it is true. But the opposition reflects the technology’s reputation. ai bosses have spent years warning of a looming job-pocalypse and the danger that an AI -engineered super-virus will make humans extinct. Opponents of data centres variously believe they are shielding the environment, protecting jobs and saving the species—and they are not entirely wrong.
Yet this backlash is itself dangerous. AI promises to change the world for the better, much as electricity or the steam engine did. Not long ago, the era-defining problem for the rich world was stagnant economic growth and the populism it unleashed. Now it has a technology that could power a surge in productivity and incomes, help find cures for untreatable diseases and improve everything from education to green tech.
All this could be lost if countries starve the technology of computing power or regulate it into uselessness. Look at m RNA vaccines research, which has been held back after a backlash during the covid-19 pandemic.
Scenarios in which some countries give in to popular rage but others forge ahead are also worrying. If America succumbs, it could cede the global ai frontier, and the attendant cyber and military capabilities, to authoritarian China. Europe and Canada are more risk-averse than America. If they choked off AI while the rest of the world kept pushing forward, their losses could be unrecoverable. More than two centuries after the Industrial Revolution, few countries have managed to catch up with the first movers.
So the stakes are high. Can governments do anything about it? Grand proclamations about the shape of a “social contract” for a post- AI world are good fodder for blog posts but offer little help today. Besides, the unknowns are still large enough to make the exercise almost futile.
Better to be incremental. While China’s economy was growing by 10% a year in the 1980s—faster than all but the most extreme forecasts for AI -driven growth—the mantra of its leader Deng Xiaoping was “crossing the river by feeling the stones”: pushing forward iteratively, planning for problems but staying flexible. Deftly handling the AI age will take a similar spirit.
To that end, here are four pointers for politicians and AI companies looking for policies. First, spread the benefits of AI as widely as possible. Blockers need to be shown that their local area will benefit if they get out of the way. Wisely, data-centre firms are beginning to offer funding to nearby towns. Gradually, this approach needs to be broadened to society at large, with mechanisms showing people that they have an economic stake in AI ‘sprogress, and will be helped to adapt to disruption through policies such as wage insurance. Only a shared sense of prosperity can temper the toxic who-wins/who-loses politics that emerged in the era of globalisation.
Second, regulate hard when interventions are needed. The hair-raising prospect of AI -enabled cyber-attacks or bioterrorism is still not taken as seriously as it ought to be. Tackling those issues and others is essential in itself, but it would also weaken arguments to ban or hobble AI indiscriminately. Ideally, these efforts would involve international co-operation.
Third, measure everything. The common view that AI is already leading to lay-offs and raising electricity bills is probably wrong. But without better statistics it is hard to be sure. Data centres must contend with viral worries over water usage, a confected issue. (Modern ones drink up no more than other industries, and much less in total than America’s golf courses.) Facts won’t cure misinformation, but their absence worsens it. Britain’s AI Security Institute and new AI Economics Institute may offer models for other countries to follow.
Fourth, use AI to make the state better. It is not just the private sector that could use AI to lift productivity. Filing taxes should be a breeze; state-run health-care systems should link up data seamlessly and schools should experiment with ai -powered learning. AI may also make it easier for citizens to monitor what politicians are up to.
People are less likely to oppose a technology if it is behind their grandmother’s cancer treatment or helping their child’s education. And they are more likely to trust that the state can oversee it if they believe that government works.
Machine politics
Voters are right to take a close interest in how AI could change their lives. The future will be messy, odd and unpredictable. Persuading them that their interests are being served by disruption has become as important as making AI models better. Failure will bring out more pitchforks—and destroy vast opportunities for humanity. ■
For subscribers only: to see how we design each week’s cover, sign up to our weekly Cover Story newsletter.
논증 분석
유형: prescription
핵심 주장
Artificial Intelligence에 대한 대중의 반발이 서방 전역에서 거세지고 있으며, 이 흐름을 방치하면 AI 리더십을 China에 내줄 위험이 있다. 정부는 거창한 선언 대신 혜택의 광범위한 분배, 특정 위해 규제, 측정 강화, 공공서비스 AI 도입이라는 네 가지 점진적 처방으로 대응해야 한다.
논리구조
- 진단 — 반발의 현황: America에서 데이터센터 반대 시위가 약 1,000억 달러 규모의 프로젝트를 차단했고, 유권자의 40%가 AI 금지를 지지하는 등 대중적 적대감이 구체적 수치로 나타나고 있다.
- 진단 — 반발의 원인: AI 기업 경영진이 스스로 ‘일자리 종말’과 AI 설계 바이러스를 경고해온 탓에 데이터센터 반대론자들이 환경·일자리·종의 생존을 동시에 수호한다는 서사를 갖게 됐다.
- 논거 — 반발 자체의 위험성: mRNA 백신 연구가 팬데믹 이후 반발로 지체된 것처럼, Artificial Intelligence 개발이 규제나 컴퓨팅 자원 차단으로 무력화될 경우 생산성 혁명·질병 치료·교육 개선이라는 막대한 기회를 영구 상실할 수 있다.
- 논거 — 지정학적 위험: America가 대중 압력에 굴복할 경우 사이버·군사 역량을 포함한 AI 최전선을 권위주의적 China에 양보하게 되며, 산업혁명의 역사가 보여주듯 선발주자를 추격하는 것은 수백 년이 걸린다.
- 처방 1 — 혜택의 공유: 데이터센터 기업의 지역사회 투자 방식을 사회 전반으로 확장하고, 임금보험 등을 통해 시민이 AI 성장에 실질적 이해관계를 갖도록 해야 한다.
- 처방 2·3 — 규제와 측정: AI 기반 사이버 공격·생물테러 위협을 실효적으로 규제하고, AI Economics Institute 등을 모델 삼아 AI의 경제적 영향을 정확히 측정함으로써 오정보를 줄여야 한다.
- 처방 4 — 국가 서비스에 AI 적용: 납세·의료·교육 등 공공 영역에 AI를 실제 도입해 시민이 기술의 혜택을 직접 체험하게 하고, 정부 역량에 대한 신뢰를 쌓아야 한다.
결론
AI 반발은 기술의 성숙과 함께 더욱 거세질 것이므로, 각국 정부는 Deng Xiaoping의 ‘돌을 더듬으며 강을 건너는’ 방식처럼 점진적·실용적 정책으로 대중의 신뢰를 획득해야만 인류 전체의 기회를 지킬 수 있다.