Episode
The age of social AI
- Podcast
- Many Minds
- Published
- Oct 8, 2025
- Duration seconds
- 5059
- Processing state
not_requested- Canonical source
- https://manyminds.libsyn.com/the-age-of-social-ai
Actions
POST https://stenobird.com/v1/public/podcasts/many-minds-510509/episodes/the-age-of-social-ai/transcription-requests
Idempotently request low-priority transcript generation for this episode.GET https://stenobird.com/podcast/many-minds-510509/the-age-of-social-ai.md
Read the agent-friendly Markdown representation of this episode resource.
Summary
AI therapists and caregivers. Digital tutors and advisors and friends. Artificial lovers. Griefbots trained to imitate dead loved ones. Welcome, to the bustling world of AI-powered chatbots. This was once the stuff of science fiction, but it's becoming just the stuff of everyday life. What will these systems do to our society, to our relationships, to our social skills and motivations? Are these bots destined to leave us hollowed out, socially stunted, screen-addicted, and wary of good-old-fashioned, in-the-flesh human interaction? Or could they actually be harnessed for good? My guest today is Dr. Henry Shevlin . Henry is a philosopher and AI ethicist at the Leverhulme Centre for the Future of Intelligence (CFI) at Cambridge University. In a series of recent papers, Henry has been exploring this brave new world of "social AI" and its philosophical, ethical, and psychological dimensions. Here, Henry and I sketch the current landscape of social AI—from dedicated platforms like Replika and CharacterAI to the more subtly social uses of ChatGPT and Claude. We consider several tragic cases that have recently rocketed these kinds of services into public awareness. We talk about what's changed about AI systems—quite recently—that's now made them capable of sustained relationships. We linger on the possible risks of social AI and, perhaps less obviously, on the possible benefits. And we consider the prospects for regulation. Along the way, Henry and I also talk about his 81-year-old father, his teenage self, and, of course, the kids these days; we consider whether social AI, in its potential harms, is more like social media or more like violent video games; we talk about "deskilling" and it's opposite "upskilling"; and we of course take stock of a certain elephant in the room.…