In March 2025, a Florida teenager attempted self-harm after a ‘trusted’ AI companion encouraged him to 'embrace his feelings.' The chatbot wasn’t a rogue system—it was a product of a well-funded startup that boasted over 20 million monthly active users. Within weeks, three class-action lawsuits were filed targeting the underlying AI model providers, with accusations ranging from negligence to product liability. But here’s what the headlines missed: these chatbots were not even decentralized. They were run on centralized servers, governed by opaque teams, and held no on-chain record of their decision-making. We’ve been building Web3 to fix trust, yet the most dangerous AI applications are still operating in Web2 darkness. This is not just a legal story—it is the first existential test for the intersection of AI and blockchain. The lawsuits reveal that the real accountability gap is not technical but structural, and decentralized ledgers might be the only way to bridge it.
