Critics Mock Anthropic's Claims Chinese AI Labs Are Stealing Its Data
Decrypt
by Jason NelsonFebruary 23, 2026
AI-Generated Deep Dive Summary
Anthropic has accused three Chinese AI labs—DeepSeek, Moonshot, and MiniMax—of extracting millions of responses from its Claude chatbot using fraudulent accounts, allegedly to train competing systems. The company claims these actions violate export controls and compromise safety measures designed to prevent misuse by foreign entities. However, this accusation has sparked widespread skepticism and criticism on platforms like X, where users argue that Anthropic itself has faced legal battles over similar practices, including a lawsuit from Reddit for scraping data without permission.
The争议 revolves around the use of "distillation," an AI training method where smaller models learn from the outputs of larger ones. While distillation can be legitimate, Anthropic alleges that these Chinese labs used it to create competing systems without proper authorization. The company warns that such activities could allow foreign entities to bypass safeguards and integrate advanced AI capabilities into military, intelligence, or surveillance systems.
Critics point out the hypocrisy in Anthropic's stance, as its own models are trained on publicly available data, including content scraped from platforms like Reddit. This raises broader questions about intellectual property, copyright, and fair use in AI development. On X, users mocked Anthropic for attempting to restrict others from using AI while engaging in similar practices itself.
The issue also highlights the challenges of regulating AI training practices across borders, particularly as export controls become a focal point in U.S.-China relations. Anthropic has called for coordinated action among industry players, cloud providers, and policymakers to address these threats. However, critics argue that the company's own history of scraping data undermines its credibility in advocating for stricter controls.
For readers interested in crypto, this story ties into ongoing debates about data ownership, privacy, and decentralized technologies. The use of distillation attacks and unauthorized data extraction raises questions about how AI models are trained and whether such practices align with blockchain principles of transparency and fairness. As AI becomes more integrated into Web3 projects, the ethical and legal implications of data usage will likely become even more critical.
Verticals
cryptoweb3
Originally published on Decrypt on 2/23/2026