May 16, In 2023, Sam Altman appeared before a subcommittee of the Senate Judiciary. The hearing was titled “Supervision of AI.” The conference was a love show, with Altman and the Senators celebrating what Altman calls a “print publisher” and acknowledging that the United States needs strong laws to avoid its pitfalls. “We believe that the government’s regulatory intervention is crucial to mitigating the risks of increasingly powerful models,” he said. “Every word the lawmakers have continued to say about Altman because he has a look at how smart laws allow AI to flourish, but only under solid guidance, both legislators and AI builders thought it was crucial at the time. Altman speaks to the industry, which widely shares his attitude. The call for battle is “regulate us!”
Two years later, on May 8 this year, Ultraman returned to another group of senators. The Senator and Artman are still singing the same tune, but one person is withdrawing from another playlist. The hearing was called “winning the AI competition.” In DC, the term “supervision” has failed, and AI discourse is no exception. Committee Chairman Ted Cruz did not advocate for external agencies to examine AI models to assess risks or remind them when interacting with AI, but instead reminded governments not only to promote the path to innovation but also to remove barriers like “overregulation.” Ultraman gets along with him. His message is no longer “regulating me”, but “investing in me”. He said overregulation – like the rules passed by the EU or a bill recently vetoed in California would be “disastrous”. “We need space to innovate and grow rapidly,” he said. He confirmed that safety guardrails may be necessary, but they need to involve “not slow our wise regulations.”
what happened? On the one hand, the moment of panic has passed after everyone is frightened by Chatgpt, and it is obvious that Congress will act soon. But the biggest development is that Donald Trump retakes the White House and hits the brakes with a subtle promotional tone of the Biden administration. The Trump doctrine of AI regulations seems to be similar to the teachings of Trump supporter Marc Andreessen, who declared in his tech optimist manifesto that AI regulations are actually a form of murder because “any AI slowdown will harm life.” Vice President JD Vance made these priorities clear at an international gathering in Paris in February. “I’m not here… talking about AI security, which is the title of the conference a few years ago,” he said. “We believe that overregulation of the AI sector could kill the transformation industry when it takes off, and we will do everything we can to encourage growth-driven AI policies.” Later, the government released an AI action plan “to enhance the U.S. as an AI powerhouse and prevent unnecessary heavy demands from hindering innovation in the private sector.”
Two enemies appeared in this movement. First, the EU has adopted a regulatory scheme that requires transparency and accountability for large AI companies. The White House despises this approach as much as those who set up AI businesses in the United States.
But the biggest bogey is China. The prospect of beating the People’s Republic of America in the “AI Competition” is so unthinkable that regulations must be put aside, or done with what Altman and Cruz call “light touch”. Some of these reasonings come from a theory called “hard takeoff” that argues that AI models can reach a turning point where lightning self-improvement can cause a stunning spiral of dazzling superpowers, also known as AGI. “If you get there first, you can’t catch you, I won’t catch you,” said former Google CEO Eric Schmidt. He is the competitor’s “you” (Schmidt has been talking about China’s position as a leader in open source.) Schmidt is one of Schmidt’s loudest voices. However, the White House may not be as interested in the singularity as it is in classical economic competition.
Worrying that China is leading in AI is a key driver of US current policy, safety is damn. Political parties even oppose countries, trying to fill the vacuum of inaction with their own laws. The House just passed tax contributions, the version of the giant Megabier of Medicaid cuts includes a moratorium on any state-level AI legislation ten years. As far as AI progress is concerned, it’s like eternity. (The rule will not survive the Senate, experts say, but it should be noted that nearly every Republican in the House voted for it.)
