Chinese content removals, local AI groups, OpenAI's creator retreat and Flock's council guide show institutions shaping public participation through different channels.
Artificial Intelligence··Midday
A platform campaign becomes standing governance
China's National Radio and Television Administration says online audiovisual platforms removed more than 13,300 AI-made videos that distorted classic films and television works in July and acted against more than 30 accounts. Global Times reports that the effort began as a special enforcement campaign in January 2026 and became a standing governance framework on February 1. Targets include altered versions of adaptations of the four great classical novels, historical and revolutionary productions, and works about exemplary figures. Platforms must also remove animation deemed to promote cults. This channel determines which AI-altered material remains in public circulation through platform enforcement. The report, however, names no participating platforms, describes no appeal route and gives no breakdown between material found through automated detection and material found through reports. The published totals show the scale of action in videos and accounts, while leaving the platform identities and the user's route for challenging a decision unspecified.[1]
Community lists and a company retreat
Rest of World reports that the AI Collective, founded in Silicon Valley in 2023, has grown to more than 200 chapters in 50 countries and over 200,000 members. Nairobi has 3,000 members and Bengaluru more than 10,000; AI Salon has chapters in Bengaluru and Lagos. About 20 people attended one AI Salon meeting in San Francisco. AI Collective chief operating officer Catherine McMillan says the aim is to build highly local communities, while Utrecht University's Payal Arora says inclusive communities cannot simply globalize Silicon Valley's conversations. The report measures neither meeting subjects, repeat attendance nor session languages. OpenAI, meanwhile, held its first content-creator retreat, Summer Club, at a luxury resort in upstate New York. The program included ChatGPT Work workshops, dinners and beekeeping. Attendees drew social-media criticism framed around data centers' environmental effects, a large planned data-center investment in Ohio and a Defense Department contract. An OpenAI spokesperson said creators matter to how people learn about its products. The report names no attendees and does not state whether a disclosure requirement applied.[2], [3]
The presentation before a local decision
A Flock Safety guide published by 404 Media shows another participation channel before automated license-plate reader systems reach city councils. Supplied to the outlet by an anonymous activist, the document advises police departments to own the narrative before public opposition forms and to reframe the discussion instead of arguing over whether the systems constitute mass surveillance. It recommends prepared presentations for resident concerns, early briefings for city managers and meetings with council members before votes. It also advises framing cost around unresolved crime rather than camera prices and credits a Richmond program with 274 arrests and 259 vehicle recoveries over two and a half years. Flock Safety did not immediately respond to requests for comment. Together, the four reports show different institutions structuring public participation: platforms remove content, communities convene local groups, a company brings creators into product workshops, and a vendor guide shapes the presentation before a local vote. Each channel establishes a different arrangement for who participates and what information enters view.[4], [1], [2], [3]
Related columns
For more information on this topic, you can read the related columns.
