Coinbase Seeks Access to Anthropic's Powerful AI Model for Cybersecurity

Crypto exchanges handling massive value in digital assets are increasingly concerned about persistent threats such as hacks, phishing, smart contract exploits, and AI-augmented attacks.

Anthropic's Mythos AI model has exceptional cybersecurity capabilities, including autonomously discovering thousands of zero-day vulnerabilities. This makes it an attractive tool for crypto firms like Coinbase to pentest their infrastructure, wallets, smart contracts, and custody systems at unprecedented scale and speed.

The access to Mythos is tightly controlled under Project Glasswing, with mitigations like monitoring to prevent misuse. If approved, the integration of Mythos into Coinbase's security stack could enhance operational resilience amid rising AI threats.