Microsoft's chief scientist warns that Donald Trump's proposed ban on state-level guardrails on artificial intelligence will slow the development of frontier technology rather than accelerate it.
Dr. Eric Horvitz, former technology advisor to Joe Biden, said the ban on regulations could be “controlled” and “conflicts on us” and “can be at odds with making good progress by not only moving forward with science but translating it into practice.”
The Trump administration proposed a 10-year ban on US states that creates “laws or regulations that limit, limit or regulate artificial intelligence models, artificial intelligence systems, or automated decision systems.”
It is driven in part by the White House fear that China can win the competition for human-level AI, but there is also pressure from high-tech investors like Andreessen Horowitz, an early Facebook investor. Its co-founder Trump donor Mark Andreessen said earlier this month that the US was dependent on two competitions with China for AI hegemony. US Vice President JD Vance recently said: [China] Don't pause? Then we find ourselves… enslaved [China]- Mediation AI. ”
Horvitz said he was already concerned about its use “AI is being used for misinformation and inappropriate persuasion” and “for malicious activities in biological hazard spaces, etc.”
Horvitz's pro-adjusting comments have emerged despite reports that Microsoft is part of Silicon Valley lobbying on Google, Meta and Amazon to help ban AI for the next decade included in Trump's budget bill passing through Congress.
Microsoft is part of a lobbying effort to encourage U.S. senators to enact a decade-long moratorium on individual states that will introduce their own efforts last week, the Financial Times reported last week. The ban is written in Trump's “Big Beautiful Bill” that should be passed by Independence Day on July 4th.
“It's as a scientist that you communicate with government agencies, especially those that may have made unregulated statements,” Horvitz said at a conference of the Association for the Artificial Intelligence Advancement on Monday. [that] This will hinder us.
“Guidance, regulations… reliability management is part of moving forward with the field and in many ways it's about making the field faster.”
Speaking at the same seminar, Stuart Russell, a professor of computer science at the University of California, Berkeley, said:
After the newsletter promotion
The obvious inconsistency between Microsoft's chief scientists and reporting company lobbying is in the growing fear that unregulated AI development can pose catastrophic risks to humanity and is driven by companies that prioritize short-term profits.
Microsoft has invested $14 billion (£10 billion) in Openai, the developer of ChatGpt. Its CEO Sam Altman said this week, “I think five or ten years from now, there's a great human robot, and I'm just walking down the street and doing something.
Predictions for when to reach human-level artificial general information (AGI) vary over several years to decades. Meta Chief Scientist Jan Lekun said AGI could be decades away, but last week's boss Mark Zuckerberg announced a $15 billion investment to achieve “ultra intelligence.”
Microsoft declined to comment.
