Image via Google Research
Monday, September 15, 2025
Google's privacy breakthrough sparks publisher lawsuit
Google dropped VaultGemma, the largest open differentially private LLM with fresh scaling laws that could reshape how we think about privacy in AI. Meanwhile, Rolling Stone is suing Google over AI Overviews (yikes), claiming the summaries are tanking their traffic, while Anthropic quietly released Claude Code hooks for deterministic automation and OpenAI launched parental controls for ChatGPT. And because speed is apparently everything now, xAI just shipped Grok 4 Fast in beta with 10x faster performance for quick queries. So here's the question: if privacy-first models like VaultGemma become viable, do publishers even have a legal leg to stand on?
Image via Google Research
Top Stories
Penske Media sued Google over AI Overviews, claiming the summaries significantly reduce publisher traffic and revenue, highlighting escalating tensions between content creators and AI companies over copyright and fair compensation for training data.
Anthropic
Anthropic's Claude Code hooks enable developers to automate and control Claude's behavior at key lifecycle points through custom scripts, notifications, and permission gates, offering fine-grained control over AI agent execution in development workflows.
OpenAI
OpenAI is introducing parental controls and mental health safeguards for ChatGPT, leveraging expert guidance and reasoning models to better support vulnerable users, particularly teens, while enabling family oversight within the next month.
Testing Catalog
xAI's Grok 4 Fast delivers 10x faster responses by optimizing for speed over sophistication, targeting users seeking immediate results for routine tasks. The model's rollout to free tier and mobile platforms signals xAI's shift toward speed-differentiated AI products.
Keep Reading
Industry Voices
Enjoyed this issue?
Get daily AI intel delivered to your inbox. No fluff, just the stories that matter.