Dr. Eric Horvitz, Microsoft’s chief scientist, believes that regulation, when carefully designed, could help, not hinder, AI development.
Speaking at a major AI conference, he emphasized that smart rules can increase safety and speed up innovation.
Horvitz called on scientists to help lawmakers understand this. According to him, fears that regulation will delay progress are misplaced.
In his words:
“It’s up to us to as scientists, to communicate to government agencies, especially to people who might be making statements like “No regulation! This is going to hold us back!” “
Horvitz believes that the right approach, guidance, and oversight can strengthen the field.
His remarks come as the Trump administration proposes a 10-year ban on state-level AI laws.
Trump’s Plan
The Trump administration argues that too many state laws could create confusion and slow down innovation. This could unintentionally give other major players, like China, a competitive edge.
Supporters argue that a single national framework will help the U.S. move faster, attract investment, and beat the competition.
Take investor Marc Andreessen, an early Facebook backer and Trump donor who voiced strong support for the bill.
He claims that AI development is a “two-horse race” between the U.S. and China. Vice President JD Vance also echoed that message.
He warned that a pause in American AI development could give China the upper hand. This urgency has led to proposals that would restrict states from creating AI laws for a decade.
The ban is included in Trump’s proposed budget and is expected to move through Congress soon.
Scientists’ Warning
Despite political pressure, experts like Dr. Horvitz urge a different path. He stressed that reliable, safe AI systems require strong oversight. In his words, “Regulation, done properly, can speed us up.”
He cautioned against slogans that reject all forms of control; instead, he urged scientists to work with governments.
He believes technical experts should help lawmakers craft effective rules. Horvitz also pointed to increasing risks like misinformation, manipulative AI tools, and the misuse of AI in biology.
Without proper limits, he says, AI could be used in harmful ways.
Another Warning
At the same event, Professor Stuart Russell from the University of California, Berkeley voiced even stronger issues.
He questioned the release of technology that may pose a serious threat to humanity. Russell said some experts estimate a 10% to 30% chance that advanced AI could cause human extinction.
He argued that no other technology with such high risk would ever be allowed to progress without strict oversight.
A Mixed Message
While Horvitz supports smart regulation, Microsoft’s actions suggest a different path. According to reports, the company is part of a lobbying effort to support the 10-year state ban.
Alongside Google, Amazon, and Meta, Microsoft is urging Congress to approve the moratorium. The companies argue that 50 different state laws would create unnecessary complications.
Fred Humphries, Microsoft’s vice president of U.S. government affairs, explained the company’s position.
He said Microsoft supports a single national framework for regulating advanced AI. This would avoid confusion and promote safer development.
However, he added that states should still have authority in areas where they typically regulate. This reflects a more nuanced position.
