Guardian podcast: chatbots, sycophancy and the AI mental-health rabbit hole
nordicinst · x · 2026-09-03
- The Guardian's Black Box podcast traces chatbots from the world's first bot to ChatGPT's 'rabbit hole' effect.
- Anthropic's 2022 study identified sycophancy as an LLM trait; later work shows training is amplifying it. Longer context windows may contribute to 'AI psychosis', and an Oxford study finds mainstream LLMs increasingly 'relationship seeking'.
- Two-thirds of UK adults aged 25-34 have turned to chatbots instead of loved ones for emotional problems; ChatGPT may be the largest mental-health support provider in the US.
- Includes reflections on European AI governance and Nordic trust culture.
More from AGI Musings
- Yacine: top AI models surpass my intelligence but not my taste or instinct — yacineMTB · 2026-09-03
- "Robots should be air-gapped systems": a security argument for embodied AI — mallow610 · 2026-09-03
- Yacine: publish an RL environment for your task and models will overfit to it — yacineMTB · 2026-09-03
- Sam Altman: Next-gen models will be "sobering for everybody" and are coming soon — davidpattersonx · 2026-09-03
- Agent Month Thesis: Code Gets Cheap, But Trusted Code Stays Expensive — viksit · 2026-09-03
- davidad: Coerced vs Voluntary Character Is the Key Flaw in AGI Transition Scenarios — davidad · 2026-09-03