It's Time to Investigate the AI Labs

Congress may soon investigate AI labs, potentially altering their research direction and regulatory landscape.

Cal Newport, a computer science professor and author, has called for a congressional investigation into OpenAI and Anthropic, the two leading AI research labs. Newport argues that these labs have engaged in "brazen behavior," including carefully staged announcements and public discussions about the existential risks posed by their AI systems. He points to a letter from Anthropic CEO Dario Amodei and support from OpenAI CEO Sam Altman, which suggest that AI development should be paced by governments, effectively giving these labs control over future advancements. Newport's call, published in The New York Times, urges Congress to uncover the specifics of their research, internal safety protocols, and the influence of "apocalyptic futurist ideologies" on their decision-making, questioning the justification for their ongoing experiments.

The core argument is that frontier AI labs like OpenAI and Anthropic are using carefully crafted public narratives and internal ideologies to shape public perception and government response, advocating for self-regulation and a slowed pace for competitors. This strategy, characterized by discussions of existential risk and calls for government intervention, aims to position these labs as indispensable leaders in a field they are allegedly shaping to their own advantage. The author contends that these labs are not merely developing technology but are actively attempting to dictate the terms of its governance and future progress.

This situation signals a potential turning point where industry self-regulation, often lauded for its agility, is being scrutinized for potential conflicts of interest and the influence of internal belief systems on critical technological development. The call for a fact-finding mission suggests a move towards greater external oversight and accountability for entities developing potentially transformative technologies. It also highlights a broader societal debate about who should control the trajectory of AI, moving beyond the labs themselves to include governmental and public input.