The News
The debate around the contours of an acceptable federal AI safety bill shifted again after a draft of a proposal Anthropic has been pitching to lawmakers on Capitol Hill leaked late last week.
Anthropic, long seen as the safety-promoting frontier lab, found itself on the defensive, with one of its top lobbyists posting about tweaks it plans to make to the draft that more accurately reflect the dizzying pace of AI progress and the company’s stance on pre-empting state laws.
That draft bill is a legislative version of an AI safety framework the company has been pitching for months, a company spokesperson told Semafor. However, AI safety advocates are arguing it doesn’t go nearly far enough, unfairly pre-empts state laws, and accepts too much possible risk.
OpenAI quickly saw an opportunity to distinguish itself from its rival, with Chief Global Affairs Officer Chris Lehane writing in a post Monday implicitly directed at Anthropic: “Policy decisions must ultimately be made through our democratic processes with our elected officials leading. … No one company (including OpenAI) should write the rules alone for a technology with such profound implications for our economy, society and national security.”
Ashley’s view
Over the past few years, Anthropic has embraced regulation in a way that other AI companies haven’t. As the frightening possibilities of its advanced AI have emerged — and as its fight with the US government has brewed — the company has started to seek regulation through methods that have promoted heated debate within the federal government (like pitching the terms that would allow officials to block the release of certain models).
OpenAI has generally been a savvier, subtler Washington player, wanting to be seen as ready to endorse compromise federal AI bills written by lawmakers, as it pursues a strategy across the states of encouraging similar frameworks it finds acceptable.