Paul Novosad says AI writing creates a trust-eroding lemons problem
paulnovosad · x · 2026-07-23
Paul Novosad argues that AI writing creates a “lemons problem”: people who use AI responsibly get mixed up with low-effort or deceptive AI-generated prose.
- He estimates that only about 1% of AI writing is genuinely useful help for someone with good ideas.
- In the other 99%, AI often masks weak thinking or lazy effort.
- Because readers learn to distrust the style, good AI-assisted writing can get penalized too.
- His takeaway: AI writing is safest inside teams with explicit norms and high trust, and should be labeled when it is heavily AI-assisted.
More from AGI Musings
- Sycophancy can stop an AI from disproving your conjecture — joshwhiton · 2026-07-23
- Matt Yglesias debates AI’s labor-market impact on a new podcast episode — binarybits · 2026-07-23
- Agentic AI is moving from assistant to systems that can do part of the work — RiciglianoMarceloE · 2026-07-23
- A long post argues AI alignment should maximize human autonomy through truth — GlenBradley · 2026-07-23
- Research explores antibodies that may help protect people from cancer — EricTopol · 2026-07-23
- Allison Duettmann imagines emulated time and many lives beyond one human lifespan — juanbenet · 2026-07-23