In the heart of tech country, voters turn on AI
Por Tyler Katzenberger — POLITICO – TOP Stories
SACRAMENTO, California — Even Silicon Valley’s backyard is bracing for a dark AI future.
California voters are overwhelmingly pessimistic about AI and believe the rapidly advancing technology poses a serious risk to humanity, according to a new poll by POLITICO and its partners. The anxiety in the hub of the global tech industry cuts across partisan and demographic lines, with voters of different ages and party affiliations expressing remarkably similar concerns about where the technology is headed.
“The drumbeat of negativism is really quite clear,” said Jack Citrin, a University of California, Berkeley, political science professor and co-director of the UC Berkeley Citrin Center for Public Opinion Research-POLITICO poll.
The results are a striking rebuke of the technology from a state that, more than any other, stands to benefit from the industry’s continued growth. More than three quarters of registered voters (77%) said advanced AI is likely a threat to humans, with 44% saying it is “very likely” a serious risk.
The anxiety in California, ground zero of the AI boom, mirrors rising concern across the country. A recent survey from POLITICO and the independent London-based pollster Public First found 63% of Americans said advanced AI poses at least a “moderate risk” of destroying humanity.
The darkening outlook lands as industry leaders and politicians debate whether AI needs tighter safety rules, following dire warnings from the highest levels of industry that unchecked innovation threatens potentially disastrous consequences for humans. Anthropic CEO Dario Amodei, OpenAI CEO Sam Altman and Elon Musk recently called for slowing the pace of AI development, while President Donald Trump — though announcing plans over the weekend to create an “AI Force” — has dismissed fears about the technology as a “hoax.”
Across the country, Democratic and Republican politicians are pushing for stronger controls.On Friday, California Gov. Gavin Newsom signed an executive order accelerating the implementation of new AI oversight standards and tasked state agencies with exploring whether a mandatory “kill switch” is needed for the most advanced AI models.
Newsom, a likely presidential candidate, said his executive action was driven in part by federal inaction. His office hasn’t ruled out calling a special session on the issue, as the governor floated last week in an interview with POLITICO.
California voters appear to be siding with more regulation on the industry.
A whopping 84% of those polled backed mandatory, independent safety testing before companies publicly release advanced AI models. And 59% said they supported adopting stronger AI rules to improve safety and fairness, even if it slows innovation and makes U.S. companies less competitive with global rivals like China.
Those views were consistent with a concurrent survey of POLITICO’s audience of key political and policy influencers in the state, including political staffers, lobbyists, policy advisers and others deeply versed in California’s political landscape. Among the roughly 400 influencers surveyed, 57% said they were more pessimistic than optimistic about AI, and 78% supported tougher AI safety rules at the risk of putting U.S. firms at a global disadvantage. That’s despite influencers indicating they use AI at work more often than the general voting public.
“Californians see AI is making them more productive, but that’s why they’re becoming more pessimistic about its broader effects,” said Jon Cohen, founder and CEO of TrueDot, whose firm fielded the poll. “There’s a clear tension.”
Still, according to the poll, AI is nowhere near a top voting issue. Just 3% of voters named data centers and AI as the “most important” issue facing California, compared to the plurality of 34% who chose “cost of living” as their greatest concern.
The voters surveyed also saw some benefits for AI in the health care industry, in accelerating their own productivity at work and in spurring economic growth. About half said they support creating a national fund that owns shares in AI giants so the public can benefit from future profits — an idea Newsom and other politicians have suggested.
These data come from parallel surveys of California voters and policy influencers, fielded by TrueDot, the AI-accelerated research platform, in collaboration with the Citrin Center at UC Berkeley and POLITICO. Interviews for the voter survey were conducted online in English and Spanish from Sept. 8-14, among a sample of 2,418 registered voters, including 1,983 likely voters, selected at random through Political Data Inc. and Verasight, with supplemental online panel recruitment. Voter results were weighted to California registered-voter benchmarks for age, educational attainment, race and ethnicity, gender, and geography, with additional calibration to 2024 presidential vote by region.
From Sept. 11-16, a parallel study was conducted in partnership with POLITICO among its audience of key political and policy influencers in California. The audience was defined based on job title and organizational affiliation and included state and local government employees, political staffers, lobbyists, policy advisers, consultants, business decision-makers and subject matter experts.
The margin of error is plus or minus 2.7% for the registered voter survey, plus or minus 3% for likely voters and plus or minus 4.5% for the influencers.
Fonte: POLITICO – TOP Stories