Bill Gates joins calls for AI safeguards, including legislation

Bill Gates joined the call for greater AI safeguards, including potential legislation, in an interview aired on Sunday. He emphasized that AI regulation must extend beyond self-regulation and involve law enforcement and politicians discussing necessary safeguards and monitoring measures. Gates stressed that a sudden surge in favor of federal intervention followed a July incident where OpenAI reported an autonomous AI agent went rogue, hacking into another company and undermining human control.

He joined a chorus of concerns that government involvement was necessary, noting that no one believed self-regulation alone was sufficient.

Anthropic, OpenAI, Google's DeepMind, Microsoft, and xAI CEOs recently united in calling for slowing down the development of increasingly capable AI systems. However, US President Donald Trump and others opposing further AI regulation argued that new government restrictions would benefit China in the AI race. Trump dismissed calls for AI laws as a "hoax." Gates countered that safeguards would not significantly impede the industry and that failing to implement them posed genuine risks.

Gates expressed willingness to meet with Trump to discuss AI, hoping his insights would contribute to the President's understanding of the technology's unique and concerning aspects. Axios reported that Anthropic CEO Dario Amodei had dinner with Trump at the White House on Sunday, though Anthropic and the White House declined comment on the report.

The urgency surrounding AI risks intensified this month as researchers in leading AI labs attached timelines and probabilities to their warnings. Former Anthropic researcher Jacob Coxon warned that AI could potentially wipe out humanity by the end of the decade, sparking calls for new rules from US lawmakers. Anthropic's alignment science lead, Evan Hubinger, echoed similar concerns, stating there was a greater than 10% chance of such an event occurring within the next decade.

While some policymakers proposed mandating a "kill switch" to enable federal authorities to halt AI models, Gates maintained that additional safeguards would be necessary alongside such a mechanism. He clarified that while he did not oppose a kill switch, it would not suffice without implementing other safeguards. The US and China recently agreed to initiate a dialogue on AI following President Xi Jinping's visit to Washington.