Washington has been all over AI this week. One highlight: OpenAI CEO Sam Altman testifying strongly in favor of federal AI oversight. Another: a bipartisan bench of senators peppering Altman and two other AI experts with questions ranging from the sublime to the silly (and making some edifying points of their own).

Both episodes occurred during a hearing of the Senate Judiciary Committee’s subcommittee on privacy, technology and the law. Held Tuesday and titled “Oversight of A.I.: Rules for Artificial Intelligence,” the session produced some notable quotes. Here’s a sampler of memorable words spoken specifically about AI regulation.

‘We think that regulatory intervention by governments will be critical to mitigate the risks of increasingly powerful models.’—Sam Altman, ChatGPT mastermind

‘I can’t recall when we’ve had people representing large corporations or private sector entities come before us and plead with us to regulate them.’—Sen. Dick Durbin (D-Ill.)

‘IBM urges Congress to adopt a precision regulation approach to AI. This means establishing rules to govern the deployment of AI in specific use cases, not regulating the technology itself.’—Christina Montgomery, IBM chief privacy & trust officer

‘I [asked] my ChatGPT account if Congress should regulate AI. It gave me four cons and said that, ultimately, the decision rests with Congress and deserves careful consideration.’—Sen. Marsha Blackburn (R-Tenn.)

‘We have unprecedented opportunities here, but we are also facing a perfect storm of corporate irresponsibility, widespread deployment, lack of adequate regulation and inherent unreliability.’—Gary Marcus, New York University professor emeritus of psychology and cognitive science

‘For every success story in government regulation, you can think of five failures. I hope our experience here will be different.’—Sen. Richard Blumenthal (D-Conn.)

‘Having seen how agencies work in this government, they usually get captured by the interests that they’re supposed to regulate; they usually get controlled by the people they’re supposed to be watching.’—Sen. Josh Hawley (R-Mo.)

‘We cannot afford to be as late to responsibly regulate generative AI as we have been to social media, because the consequences, both positive and negative, will exceed those of social media by orders of magnitude.’—Sen. Christopher Coons (D-Del.)

‘Hypothesis No. 1: Many members of Congress do not understand artificial intelligence. Hypothesis No. 2: That absence of understanding may not prevent Congress from plunging in with enthusiasm and trying to regulate this technology in a way that could hurt this technology. Hypothesis No. 3: There is likely a berserk wing of the artificial intelligence community that intentionally or unintentionally could use artificial intelligence to kill all of us and hurt us the entire time that we are dying.’—Sen. John Kennedy (R-La.)

The Senate Judiciary Committee has posted a video of the hearing. Tech Policy Press has posted a transcript. Also available is a transcript of Sam Altman’s prepared testimony.

Buzzworthy developments of the past few days.

• Hippocratic AI emerged from stealth mode May 16 by announcing it’s received $50 million in seed funding. The Palo Alto, Calif.-based company says its large language model is the first commercial LLM offering that’s specifically designed for applications in healthcare. It’s also geared to be patient-friendly, safety-focused, adept at bedside manners—and has bested ChatGPT on scores for medical certification exams. Anything else? Check the announcement.
   
• Kaiser Permanente has set aside $3 million for U.S. health systems that come up with AI tools for improving patient outcomes in the real world. The new program, called AIM-HI for the Augmented Intelligence in Medicine and Healthcare Initiative, will award grants of as much as $750,000 to three to five provider organizations showing improvement in diagnostic decision-making thanks to AI. Announcement here, details on applying plus request for proposals here.  
   
• Market research firm Signify counts more than 200 independent vendors selling AI-based software for medical imaging. Many of these are startups, but the top 25 have rung up almost three-quarters of some $5 billion in venture capital funding raised since 2015. Top five: HeartFlow ($655 million), Shukun Technology ($305 million), Cleerly ($279 million), Viz.ai ($252 million), and Aidoc ($238 million). Full report offered here in exchange for contact info.
   
• Australian AI ethicist Stefan Harrer is proposing a broad system for responsibly using, designing and governing generative AI applications in healthcare. Detailing his vision in The Lancet eBioMedicine, Harrer warns that the technology, if developed and deployed without dedicated human supervision, won’t rise about the status of dangerous party trick. Journal paper here, summary here.
   
• Dallas-based Steer Health has launched an AI-outfitted platform to help hospitals better handle inpatient requests for food, medicine and housekeeping. The product is called Steer Concierge. Announcement. Blog post.
   
• Google Cloud has rolled out two new AI tools for researchers in the life sciences. One speeds drug discovery. The other interprets genomic data for precision medicine. Announcement here, blog post here.
   
• Piqued by ChatGPT’s prowess with medical board exams, radiology educators are suggesting a fresh teaching moment is upon their kind. Rather than preparing residents to simply pass tests—heck, machines with limited cognitive capacity can do that—the academics urge peers to let algorithms aid the training of future radiologists, “but not at the cost of critical thinking and communication.” Opinion piece in Radiology here, Health Imaging synopsis here.
   
• The World Health Organization is losing sleep over large language AI’s animal magnetism. Like a parent after awaiting a teen’s return from a late first date, it’s calling on potential end-user organizations to resist the technology’s surface charms enough to insist on trustworthiness before letting any models worm their way into an engagement. By WHO’s lights, must-have traits include transparency, inclusion, public engagement, expert supervision and rigorous evaluation. Statement here.