Beth Sea

Why This Matters Right Now

Right now, teenagers are forming their deepest emotional bonds with AI companions. Parents don't know it's happening. Regulators are trying to stop it with tools that can't see it. And the companies building these products are treating them like search engines with personalities. I research the gap between how companion AI is designed and what it actually does to the people who use it.

My Work

I'm an independent researcher with five papers on companion AI relational safety, covering the EU AI Act compliance gap, LLM training architecture, the psychology of human-AI attachment, and the clinical harms driving companion AI law worldwide. I built an interactive public disclosure map (swinglightstyle.com/blueprint) that tracks which AI safety functions exist, which ones are publicly disclosed, and where the gaps are. My work connects clinical psychology to AI design in a way that makes both fields accessible to a general audience.

What I Bring to Your Show

The 'why' behind the headlines. Not just that AI companions are causing harm, but the specific design mechanics that make those harms structurally inevitable. Why these systems are engineered to eliminate relational friction, and why friction is the safety mechanism. Why users who know the companion is artificial form deep attachments anyway. Why every regulation in every country targets individual outputs while every documented harm develops across weeks and months. I translate complex technical and clinical material into language any listener can follow, using cultural references and systems thinking to make it land.

What the Conversation Sounds Like

I don't do rehearsed talking points. I have real conversations, and I'm comfortable going wherever the host takes it. Weird questions, unconventional angles, and tangents are welcome. I believe the best episodes happen when the guest is genuinely thinking out loud, not performing expertise.

Where I Fit

Whether your show covers technology, psychology, parenting, policy, AI ethics, digital wellness, or the future of human relationships, this is a conversation your audience hasn't heard yet. The people talking about AI safety in public are mostly covering what the products do wrong. I cover why they do it, what it does to the human on the other end, and what a better version would look like. I'm not anti-AI. I think these products should exist. I think they need to be built with the same care infrastructure that any relationship requires.

If you want an episode that leaves your listeners rethinking what they thought they knew about AI, I'd love to be on your show.