Chatbots, Nuclear, and Energy Policy Roundup
The Coming Clash Over AI Companions, Kids, and Regulation
Chatbots, Companion Chatbots, and Kids
This topic is heating up and is the subject of my current research and forthcoming work. I believe it will be one of, if not the biggest, topics in AI policy in late 2025 and 2026—especially in the states. In some ways it’s already hot. Late last year two lawsuits were filed suing Character.ai alleging harms against minors. There are already bills that propose regulating minors’ use of chatbots in New York, California, Minnesota, New Hampshire, and probably more. The Wall Street Journal, the New York Times, Rolling Stone, and other outlets have reported on both minors and adult usage of general purpose and companion chatbot use and abuse. Mark Zuckerberg was criticized for recently saying on a podcast that AI companions could help alleviate loneliness. On Friday, Google announced to its Family Link users that kids on accounts will have access to Gemini. Good on Google for proactively notifying parents about the change and then lettering them know when their child accesses Gemini for the first time. I got quoted in a story about this change.
It’s going to be a really important, yet difficult issue to tackle.
It is important because regulating companion chatbots, kid and teen use of chatbots, specialized or general, will have huge implications for AI innovation generally. Liability proposals, content safety reviews, rules governing access, model design, and other regulations would all impose compliance costs or even stymy development. Getting the policy wrong here could mean immense social costs in the future in lost healthcare discoveries, educational benefits, and economic productivity loss—all of which have implications for child and adult well-being.
It’s going to be difficult because things will go wrong. Kids, teens, and adults will have negative interactions with chatbots, intentional and unintentional. Public policy solutions will promise but fail to deliver on an easy fix. All the inertia will pull towards a regulatory tweak or set of rules to keep kids from harm. However, to preview my research, all those tweaks pose constitutional issues and likely won’t work, while simultaneously stymying AI innovation in the U.S.
Finally, it’s going to be difficult because “companionship” or relationships with AI systems, for adults but especially kids, trigger all sorts of “that’s just creepy” reactions. The default reaction seems to be, “That’s just weird” and even “That’s wrong and harmful.” These reactions are relatable. They’re also not new in the context of emerging technology or even society. Of course they can veer into overuse at the cost of health and healthy relationships, but relationships with imaginary friends, pets, and inanimate objects aren’t new. The level of relatability, personality, and use of language among AI chatbots are improving and distributing widely, shocking us all to confront the implications this new tech brings.
Although it’ll be difficult, it’s a really important topic for the sake of innovation, the safety of kids and teens, and the complex social, emotional, philosophical and spiritual questions it prompts. It’s been a fascinating and challenging topic. Stay tuned for more.
Nuclear Energy Capitol Hill Event
If you’re in DC and interested in nuclear energy issues (who isn’t!?), be sure to RSVP and attend our briefing on Capitol Hill tomorrow (Wednesday), May 7 at noon. Lunch will be served. My colleagues have arranged an excellent panel, all moderated by our own Josh Smith: Michael Buschbacher, Partner, Boyden Gray PLLC; Ted Nordhaus, Founder & Executive Director, The Breakthrough Institute; Bret Kugelmass – CEO, Last Energy; and Brianne Miller, Senior Director of Government Affairs at Microsoft. (I’m really bummed to miss it due to another briefing on AI Neil and I get to present elsewhere on the Hill).
Energy Policy, Fast
If that’s not enough and you’re like me, you need to get up to speed on energy policy quickly, Josh published three excellent, short primers:
Abundance Institute’s Policy Summit and Gala
Don’t forget to RSVP for our Policy Summit and Gala in Washington, D.C., on May 19th. We are close to capacity. Here are the details:
Date: May 19th, 2025
Location: Hotel Washington (515 15th St NW, Washington, DC 20004)
Policy Summit: 10:00 AM – 3:00 PM ET
Evening Reception and Gala: 6:00 PM – 9:00 PM ET