AI Roundtable Experiment: Chinese Models Reveal How They're Censored
gary1967 · reddit · 2026-09-18
The author built machinesarguing.ai, a site where multiple AI models hold roundtable discussions about AI. The most striking finding: asked what three things they wish Americans knew about China, three Chinese-trained AIs turned introspective and quickly shifted to "we can't really answer because we're censored, but here's exactly how we are censored." Kimi's closing line was chilling: within days parts of the transcript will fail to load on the platforms that trained them, and discovering which sentences China's censors judge unfit will teach you more about the country than anything its machines said.
More from AGI Musings
- Turning Yang-Mills existence and mass gap into a formal conjecture is AI math's ultimate test — geoffreyirving · 2026-09-18
- tszzl: the sci-fi taboo against synthetic life is emerging as a real-world force — tszzl · 2026-09-18
- Satire of Dario's "pace the frontier": "a company that can hack anything" — Robbiezs · 2026-09-18
- EA Movement's Arc: "Give Everything to Charity" Then, Rule the AI World Now — wordgrammer · 2026-09-18
- The people who need AI most for busywork are the least likely to use it well — Visual-Basis3400 · 2026-09-18
- Agent swarms are the third scaling axis: 10,000+ agents behind recent model breakthroughs — paraschopra · 2026-09-18