How Musk’s toxic data empire is powering Anthropic’s ‘responsible’ AI

The reliance on scraped social media data for training responsible models highlights the critical ethical disconnect between massive data acquisition and safety standards. This revelation exposes the ongoing difficulty of vetting training sets while developers scramble to secure high-quality conversational data.