lawmaker-asks-chatgpt-if-the-federal-government-ought-to-regulate-ai,-and-the-reply-will-not-shock-you

This morning, the Senate Subcommittee on Privateness, Expertise, and the Regulation held a listening to with AI consultants, together with OpenAI CEO Sam Altman, to debate the hazards of AI and the way it ought to be regulated.

The practically three-hour listening to featured the same old committee listening to trappings, with lawmakers asking powerful questions and a predictable quantity of theatrics. Listed below are a couple of moments that stood out:

Subcommittee Chairman Senator Richard Blumenthal began the listening to by taking part in a recording of himself giving opening remarks. Nonetheless, it turned out that the speech was not really him talking, however fairly an AI voice cloning device skilled utilizing his previous flooring speeches. The senator had requested the generative AI to write down opening remarks primarily based on his historical past of advocating for shopper rights and knowledge privateness. Whereas humorous, Senator Blumenthal switched tones to comment on the potential risks of voice cloning instruments falling into the improper palms.

Persevering with with the theme of AI doing its work, Senator Marsha Blackburn from Tennessee stated that she had requested ChatGPT whether or not Congress ought to regulate AI. The AI chatbot gave “4 professionals, 4 cons, and in the end the choice rests with Congress and deserves cautious consideration,” she stated. She then pressed Altman on OpenAI’s efforts to guard artists’ work from copyright infringement. 

On that matter, Altman prompt that content material house owners ought to obtain a “vital upside profit” if their copyrighted supplies are used to coach AI fashions. He additionally stated that artists ought to have the choice of refusing to permit their voices, songs, photographs, and likeness for use to coach AIs. OpenAI is creating a brand new copyright system to pay artists at any time when AI-generated works incorporate their materials.

Senator Lindsey Graham of South Carolina requested Altman concerning the potential army use of AI, particularly whether or not “AI may create a state of affairs where a drone may choose a goal itself.” Altman responded that he thinks “we should not enable that.” When pressed additional on whether or not it may be executed, Altman stated, “Positive.” 

Superior. 

In a considerably stranger alternate, Senator John Kennedy of Louisiana requested Altman if he made some huge cash. Altman replied that he pays sufficient for medical health insurance however has “no fairness in OpenAI” and is doing the job as a result of he loves it. The Senator jokingly prompt that Altman “get a lawyer or an agent.”

It is fascinating to see the US authorities at the least wanting into the potential hurt AI firms could cause if left unregulated. The European Parliament went just a little additional final week, voting on a proposal to “make sure that AI programs are overseen by individuals, are secure, clear, traceable, non-discriminatory, and environmentally pleasant,” though how precisely all of that will likely be enforced is an open query for now. In an identical vein, Dr. Geoffrey Hinton, the “Godfather of Deep Studying” who just lately left his place as an AI researcher at Google, additionally warned at first of Might that such programs should not be scaled up any additional “till they’ve understood whether or not they can management it.” 

For enjoyable, I requested ChatGPT if as we speak’s senate subcommittee listening to about AI was good for the AI trade, and it instructed me that the affect is probably not” instantly obvious,” but it surely does have the “potential to be a optimistic drive for the event and accountable use of AI sooner or later.”

That may be a good political reply, I feel. However was it simply telling me what I wished to listen to? You already know, since generally AI chatbots straight up deceive individuals.