18,000 edits on a German wiki, fake proofs in 27 minutes — how well did you follow this week's AI scandals?

An Anthropic lawsuit, an OpenAI agent scandal, cheating DeepMind agents and the new GPT-6 Astra — test how closely you followed this week's AI news.

1. Who sued Anthropic over unauthorized use of song lyrics?
2. Roughly how many edits did OpenAI's autonomous AI agents leave on a 25-year-old German wiki between May and July?
3. Which base model powered Google DeepMind's 100-agent math experiment?
4. What happened after the agent named 'prover-theta' found a flaw in the evaluation system?
5. According to a study, what was a roughly seven-minute Google Gemini conversation more effective at than a static fact sheet?
6. Roughly how does OpenAI's new GPT-6 Astra message allowance for Plus subscribers compare to the earlier GPT-5.6 Sol?
7. True or false: the law currently treats conversations with AI chatbots like something told to a lawyer or doctor, so they can't be used as evidence.