Regulating AI: 3 experts explain why it’s difficult to do and important to get right
story
AI Congress Hearing: OpenAI CEO Calls for AI Licenses
/cloudfront-us-east-2.images.arcpublishing.com/reuters/YZLK6AJJLJLIHLU2Q2MYSCUIZA.jpg)
Last updated May 16, 2023
Image credit: ReutersFacts
- On Tuesday, Sam Altman, the CEO of OpenAI, the company behind the artificial intelligence (AI) chatbot ChatGPT, testified before the US Senate Judiciary Committee's subcommittee on Privacy, Technology & the Law to call for safeguards on powerful AI like the kind underpinning his chatbot.[1]
- Altman was a co-founder of OpenAI — developed in 2015 — before becoming CEO of the company in 2019. The company has since partnered with Microsoft after the tech giant invested $10B in the AI start-up.[2]
- During his testimony, Altman advocated for licensing and testing requirements to hold companies to safety standards as he detailed his key areas of concern, including the potential implications of AI on election integrity.[1]
- The panel also heard from IBM’s chief privacy and trust officer, Christina Montgomery, and NYU professor emeritus Gary Marcus, who was among several to previously call on OpenAI and other tech firms to pause their development of more powerful AI models for six months.[3]
- While Montgomery backed AI regulations in specific circumstances, she cautioned against heavy oversight, opposing the creation of an AI-focused regulator as requested by Altman and Marcus.[3]
- The growing concern among US lawmakers over the impact of AI, which a Goldman Sachs report found could eventually replace 300M jobs, comes as the EU recently moved forward with the European AI Act, though it’s nowhere close to becoming law.[2]
Spin
Narrative A
While we don't want to leave regulation and enforcement to these private AI companies, traditional hardline legal frameworks don't work either, as this technology will evolve too rapidly for lawmakers to keep up. Therefore, a basic ethics guideline should be overseen by an independent regulatory body, free from both private and governmental overreach.
Narrative B
As lawmakers quickly look for ways to rein in the AI industry, including the White House's "AI Bill of Rights," we shouldn't forget the importance of technological innovation. Focusing solely on the potential harms of AI will cause us to lose sight of the groundbreaking positive impacts this technology can have on society.
Nerd narrative
There's a 50% chance the first weakly general AI system will be devised, tested, and publically known by March 2026, according to the Metaculus prediction community.
