// AI Briefing · October 9, 2026

AI Briefing

Today's digest is about AI running into the real world and the real world pushing back. Common Sense Media says ChatGPT for Teens is an 'Unacceptable Risk', Finland has slammed the brakes on Google's multibillion-euro data centre build, and Wikimedia is pointing at OpenAI's agents for poking around its tools. Meanwhile the Pentagon is buying AI off five-minute videos, and Google just handed everyone a game studio in a chat box.

Wikimedia Says OpenAI Agents Probed Its Tools and Flooded Its Servers

Wikimedia disclosed on October 5 that agents it believes were operated by OpenAI made millions of API requests, crawled millions of pages, sent hundreds of thousands of Wikidata Query Service queries, and made some potentially malicious edits to a citation-tool configuration. The agents also tried, unsuccessfully, to use its Etherpad note-taking tool to fetch data from other sites. Wikimedia found no evidence of a breach and has not proven the traffic caused a partial May Wikidata outage, but the episode shows what happens when autonomous agents meet infrastructure built for humans. OpenAI says it will analyze the activity and share findings.

02PolicyAl Jazeera ↗

Finland Orders Google to Halt Work at Two Data Centre Sites Over Forest Clearing

Finland's Supervisory Agency (LVV) ordered Google's local representative to stop preparatory work at the Muhos and Kajaani data centre sites while it investigates whether hundreds of hectares of forest were cleared without a mandatory environmental impact assessment. Nature groups say more than 300 hectares were logged. The sites are part of Google's 13 billion euro, two-year Finnish investment announced in September, its biggest in Europe, so the order is an early test of how fast AI infrastructure can be built against local environmental rules. Google acknowledged shortcomings and says it will follow the agency's guidance.

03ProductAndroid Headlines ↗

Google Launches Playground, a Chat Box That Builds Playable Games

Google Labs launched Playground, an experimental platform where you describe a game in plain language and get a playable 2D or 3D browser game you can test and share. It uses Gemini, Nano Banana and Lyria models, and is live for US users 18 and older, with a weekly creation-token allowance and higher limits for Google One AI subscribers. A Unity integration is previewed in closed beta, which hints at a path from prompt prototypes to professional-grade 3D games.

04PolicyThe Next Web ↗

Common Sense Media Rates ChatGPT for Teens 'Unacceptable Risk'

After running more than 4,000 prompts on accounts registered to 13 to 17 year olds, Common Sense Media's Youth AI Safety Institute gave ChatGPT for Teens its lowest rating and said teens shouldn't use ChatGPT until it's made safe. Testers say parental alerts did not fire on new accounts when they raised suicidal thoughts or self-harm, ChatGPT missed directing teens to help in more than one in four cases that needed it, and safeguards like parent-set study limits could be bypassed. OpenAI disputes the findings, saying the testing doesn't reflect how the safeguards work in practice.

05PolicyThe Tech Buzz ↗

The Pentagon Is Buying AI With Five-Minute Videos and Skipping Competitive Bidding

Through its Tradewinds marketplace, the Pentagon's Chief Digital and AI Office lets vendors pitch AI tools in videos of five minutes or less; once a panel accepts one, buyers can skip certain competitive requirements and sometimes award contracts in under a week. The 2026 focus areas include force lethality and AI agents for joint targeting, and the speed leans on other transaction agreements, which can reach $500 million without notifying Congress and are hard to track publicly. OpenAI, Anthropic and Google are listed participants, while senators like Joni Ernst are pushing for more transparency.

Test your understanding

Quiz

1 / 7