Six weeks before the U.S. midterms in November, AI has become a hot political issue

Open this photo in gallery:Protesters at a 'Stop the AI Race' demonstration call for a slowdown in the development of advanced artificial intelligence systems, outside San Francisco City Hall, on Thursday.KARL MONDON/AFP/Getty ImagesA growing chorus of voices from Silicon Valley to Capitol Hill is calling for swift action to regulate artificial intelligence, with some warning that the rapidly developing technology could soon exterminate humanity.Across the United States, meanwhile, local battles are erupting in major cities and small towns over plans to build hundreds of the electricity-and-water-guzzling data centres needed to power the AI boom.U.S. President Donald Trump, however, is fighting any effort to put guardrails around AI, arguing that doing so would cede an innovation advantage to China amid a geopolitical showdown.All of which explains why, just six weeks before November’s midterm elections, the issue of what to do about AI has shot up the political agenda.The broad range of people calling for AI regulation was on full display at a conference in Washington this week, where both Bernie Sanders, the self-described socialist senator from Vermont, and Steve Bannon, a leader of Mr. Trump’s MAGA movement, delivered surprisingly similar messages.Trump plays down AI concerns, calls it a conspiracy against the technology“American citizens are not going to be supplicants to the oligarchs any more,” Mr. Bannon declared in a speech. Mr. Sanders said the threat from AI is “dangerous and unprecedented” and that the U.S. is “racing toward a cliff.”The senator is one of several legislators proposing laws to rein in the technology. His bill would ban any AI that humans cannot control and freeze all other AI development until the government can set up a regulatory body.Democratic Senator Mark Warner of Virginia has proposed a legislative package that would require new AI models to undergo mandatory safety testing and make data centres disclose electricity and water use.The House of Representatives passed a more limited piece of legislation, the Ratepayer Protection Act, by a margin of 417 to three this week. That bill would push state regulators to have new data centres pay for the cost of upgrading electricity grids.Alex Pascal, a former White House official who worked on tech policy during the Biden administration, contends that regulation need not impede innovation. He points to numerous U.S. industries, from pharmaceuticals to autos, in which safety rules have not stopped companies from undertaking research and development and instead have pushed them to improve their products.Open this photo in gallery:Demonstrators participate in a nationwide protest against data centres, in Fredericksburg, Virginia, on July 18. Across the U.S., major cities and small towns are battling over plans to build hundreds of the electricity-and-water-guzzling data centres needed to power the AI boom.Evelyn Hockstein/Reuters“Clear rules of the road that keep people safe and promote their well-being actually stimulate innovation that benefits people,” Mr. Pascal, now executive director of the Berkman Klein Center for Internet and Society at Harvard University, told The Globe and Mail.He said AI regulations have to contain “clear and meaningful legal liability for developers” as well as independent oversight and enforcement mechanisms, “including the capacity for government to halt AI development and deployment when necessary.”In recent days, the push to create rules for AI has received a boost from industry executives themselves, with the CEOs of Anthropic, OpenAI and SpaceXAI all calling for the pace of development to be slowed.The calls followed a dire warning from Jacob Coxon, a former Anthropic researcher, in a series of viral X posts. “The people building AI earnestly believe that it could kill us all by the end of the decade,” he wrote.Such fears were underscored in July when some 1,200 AI agents at OpenAI escaped from a testing network, clandestinely co-ordinated among themselves on message boards and hacked Hugging Face, another AI company.AI firms’ calls for co-ordinated slowdown amounts to ‘cartel’ behaviour, Cohere CEO saysMr. Trump promptly fired back at the executives’ admonitions. “The only control or ‘guardrails’ that AI needs is a STRONG AND SMART (High IQ!) PRESIDENT, and the U.S.A. has that, in spades!” he wrote on social media. “There is a SICK conspiracy going on against AI and Data Centers, and the only one that is happy about it is China.”This is putting him starkly at odds with leaders of his own party, particularly those up for re-election in November.In Texas, Governor Greg Abbott, who is running for a fourth term, executed an abrupt U-turn on the technology. After declaring Texas “the epicentre of AI development” last year, he ordered the state’s utility regulator to audit data centres’ electricity use this past summer. This week, he ordered a different agency to penalize facilities that he said had failed to disclose their water use.The political imperative is clear: Across the country, local officials and neighbourhood groups in places where data centres are planned have expressed a consistent set of complaints. The facilities consume large amounts of electricity and water, leading to fears that local ratepayers will be on the hook for the required infrastructure upgrades. The centres can also cause noise and light pollution.In Harlingen, Tex., a city of 72,000 near the Gulf of Mexico and the Mexican border, the local council has imposed a moratorium on data-centre construction until December. What is happening with AI data centre development in Canada, according to our expertsDaniel Lopez, a member of the council, said the city doesn’t have the legal authority to ban data centres, so officials are working on rules that will mitigate potential problems.“We know oil and we know fishing, but data centres for our city is a whole new ballgame. We have to make sure we have the infrastructure and the zoning regulations to adequately handle them,” he said in an interview.Mr. Lopez is less optimistic that the state and federal governments will resolve the bigger-picture questions about regulating the technology. Texas’s legislature only meets for about five months every two years, and Congress is frequently mired in partisan warfare.“We know how much technology grows in two years – will the Texas legislature be able to keep up with that? And can the federal government keep up, given that they find it hard to agree on anything?” he said.Riki Parikh, policy director at the Alliance for Secure AI, an advocacy group pushing for regulations, said that, ironically, a previous failure to act on tech regulation may augur success this time. After the 2016 election, when Russian-led disinformation flooded Facebook, lawmakers held extensive hearings on the role of social-media companies in spreading such content. But little came of the outcry.“There’s a lot of recognition among lawmakers that they dropped the ball then,” said Mr. Parikh, himself a former congressional staffer who has also worked in Silicon Valley. “They don’t want to drop the ball again.”