AI research leaders testified Tuesday before the Senate Judiciary Committee to discuss the prospects and challenges of AI technology. The general consensus was a call for swift yet considerate action to prevent the abuse of AI or to avoid hindering the industry’s progress. Experts included Dario Amodei, Anthropic co-founder; Stuart Russell, UC Berkeley; and Yoshua Bengio, a long-time AI researcher.
Amodei called for the securing of supply chains, emphasizing the creation of a testing and auditing process akin to those for vehicles and electronics. However, he highlighted that the science for establishing these measures was still developing. He expressed concern over misinformation, deepfakes, and propaganda, especially during election seasons.
Bengio stressed the importance of limiting access to large-scale AI models, ensuring alignment, and tracking who has access to the hardware required to develop these models. He underlined the need for global funding for AI safety research and cooperation among nations rather than competition. He pointed out that even smaller actors can fine-tune pre-trained models, posing a potential threat. Bengio called for the establishment of a single regulatory body per country to coordinate efforts and reduce bureaucratic slowdowns. Moreover, he suggested that social media accounts should be “restricted to actual human beings that have identified themselves, ideally in person.”
Russell advocated for an absolute right to know if one interacts with a machine or a human and proposed the outlawing of algorithms capable of making lethal decisions. He also recommended mandatory kill switches for AI systems if they hack into other computers or replicate themselves and the removal of systems that breach rules. He cited “external impact campaigns” using personalized AI as the most significant risk.
Russell stressed the considerable funding flowing into AI startups, estimated at around 10 billion dollars monthly, far exceeding the budget for basic research through the National Science Foundation. When asked about China, Russell stated that their expertise in AI has been slightly overestimated.
In conclusion, the experts urged the need for investing in basic research to base proposed testing, auditing, and enforcement schemes on rigorous science. In response, Sen. Blumenthal (D-CT) mentioned that the hearing aimed to help form a government body capable of swift action given the urgency of the situation. He concluded, “I don’t know who the Prometheus is on AI, but I know we have a lot of work to make sure that the fire here is used productively.”
First published at Techcrunch.com.
Jack McPherrin ([email protected]) is a managing editor of StoppingSocialism.com, research editor for The Heartland Institute, and a research fellow for Heartland's Socialism Research Center. He holds an MA in International Affairs from Loyola University-Chicago, and a dual BA in Economics and History from Boston College.