News Summary

Google has been utilizing Anthropic’s AI model, Claude, to improve the capabilities of its own AI, Gemini. Contractors working on Gemini are comparing its output against Claude’s, as reported by TechCrunch. This practice involves evaluating responses based on various criteria such as truthfulness and verbosity, with some outputs explicitly identifying as Claude’s. The comparison process has raised questions about ethical boundaries and intellectual property, particularly since Google’s access to Claude without explicit permission from Anthropic remains unclear. Despite Google’s significant investment in Anthropic, it has not confirmed whether it has obtained approval to access Claude for its evaluations. This development comes amid shifts in the AI market, with Claude capturing a significant portion of market share previously held by ChatGPT.