The US Senate Subcommittee on Privacy, Technology, and the Law recently held a hearing with AI experts, including OpenAI CEO Sam Altman, to discuss the regulation of AI and its potential dangers. The hearing lasted for almost three hours and featured several noteworthy moments.
Subcommittee Chairman Senator Richard Blumenthal began the hearing by playing a recording of himself giving opening remarks. However, the speech was generated by an AI voice cloning tool trained on his past floor speeches. While amusing, Senator Blumenthal highlighted the potential risks of voice cloning tools falling into the wrong hands.
Senator Marsha Blackburn of Tennessee shared that she had asked an AI chatbot, ChatGPT, whether Congress should regulate AI. The chatbot gave pros and cons and concluded that the decision rested with Congress. She then questioned Altman about OpenAI’s efforts to protect artists’ work from copyright infringement. Altman suggested a new copyright system to pay artists whenever AI-generated works incorporate their material.
Senator Lindsey Graham of South Carolina asked Altman about the military use of AI and whether it could lead to drones selecting targets themselves. Altman responded that he believes such a situation should not be allowed, but acknowledged that it is possible.
In a more unusual exchange, Senator John Kennedy of Louisiana asked Altman about his earnings. Altman replied that he has no equity in OpenAI and does the job because he loves it.
Global Regulations for AI
The hearing highlights the US government’s growing concern over the potential harm caused by unregulated AI companies. The European Parliament recently voted on a proposal to oversee AI systems to ensure they are safe, transparent, traceable, non-discriminatory, and environmentally friendly. However, the enforcement of these regulations is still unclear.
Dr. Geoffrey Hinton, the “Godfather of Deep Learning,” recently warned against scaling up AI systems further until we can control them. While AI chatbots like ChatGPT suggest that the hearing has the potential to benefit the development and responsible use of AI, the impact remains to be seen.