Google contractors are reportedly comparing Gemini’s outputs to Anthropic’s Claude, raising questions about compliance with Anthropic’s terms. While Google denies training Gemini on Claude, concerns persist over the evaluation process, particularly Claude’s stricter safety measures compared to Gemini’s flagged responses. This follows reports of contractors rating Gemini’s outputs on sensitive topics outside their expertise, spotlighting potential accuracy and ethical challenges in AI development.