Microsoft鈥檚 chief scientist, Dr. Eric Horvitz, says well-crafted regulation could accelerate AI progress rather than hinder it countering growing anti-regulation rhetoric in U.S. tech and politics.
Speaking at the Association for the Advancement of Artificial Intelligence, Horvitz urged scientists to help policymakers understand that 鈥済uidance, regulation, reliability, controls鈥 are not obstacles but enablers of faster, safer innovation. 鈥淲e should be cautious about slogans like 鈥榥o regulation鈥 鈥 done properly, it can speed us up,鈥 he said.
His remarks come amid reports that Microsoft is part of a Silicon Valley lobbying push, alongside Google, Meta, and Amazon, supporting a Trump-backed 10-year federal ban on state-level AI regulation. The ban, included in Trump鈥檚 proposed budget bill, reflects investor fears that China may outpace the U.S. in AI, and is backed by figures like venture capitalist Marc Andreessen and U.S. Vice President JD Vance.
The apparent contradiction between Horvitz鈥檚 stance and Microsoft鈥檚 lobbying efforts highlights tensions within the AI industry, where calls for responsible oversight clash with commercial concerns over fragmented state laws. Microsoft鈥檚 U.S. government affairs VP Fred Humphries defended the federal approach, saying: 鈥淲e cannot afford to wake up to a future where 50 different states have enacted 50 conflicting approaches.鈥
Meanwhile, fellow AI expert Prof. Stuart Russell of UC Berkeley issued a stark warning, questioning why society would 鈥渄eliberately allow the release of a technology which even its creators say has a 10% to 30% chance of causing human extinction.鈥
Despite the ongoing debate, Microsoft has heavily invested in AI 鈥 pouring $14 billion into OpenAI. CEO Sam Altman recently predicted humanoid robots 鈥渄oing stuff鈥 on city streets within five to 10 years. Predictions about achieving artificial general intelligence (AGI) remain varied, with Meta鈥檚 Yann LeCun estimating decades, while Mark Zuckerberg has committed $15 billion to chasing 鈥渟uperintelligence.鈥
The debate underscores the growing urgency to align innovation with responsibility before the technology outpaces oversight.
The headline and text of this article were amended because an earlier version incorrectly summarised remarks of Eric Horvitz as saying that Trump鈥檚 plan to ban US states from AI regulation 鈥渨ill hold us back鈥. To clarify: he said that AI regulation, done properly, can 鈥渟peed us up鈥. For context, more of Eric Horvitz鈥檚 remarks have been included and a comment from Microsoft has been added.