Anthropic Accuses 3 Chinese Companies of Harvesting Its Data

San Francisco-based AI start-up Anthropic recently accused three Chinese companies—DeepSeek, Moonshot, and MiniMax—of using approximately 24,000 fraudulent accounts to harvest its data for training their own chatbots, following last year's market disruption when DeepSeek released AI technologies that matched the performance of leading models. The allegations come as the global AI industry intensifies its competitive race, with companies increasingly concerned about protecting their proprietary algorithms and training data from unauthorized access. Last year, DeepSeek's emergence sent shockwaves through Silicon Valley and US financial markets, as its AI capabilities suddenly matched established players without the typical development trajectory that Western companies had followed. This incident has raised questions about the methods some Chinese tech firms may employ to rapidly advance their AI capabilities, potentially through questionable data acquisition practices rather than genuine innovation. The accusation highlights the growing tensions in the AI sector where intellectual property boundaries remain unclear and competitive advantages can be gained through various means, both legitimate and questionable.