AI model access curbs deepen security risks for crypto industry

Access restrictions on high-performance AI models are widening a security gap for much of the crypto industry, leaving many firms without the latest defensive tools as AI developers try to curb cyberattack misuse, Cointelegraph reported. Some organizations, including Coinbase and Shielded Labs, a Zcash ecosystem entity, have secured access to Anthropic’s Mitos and conducted security audits, according to the report. Many large crypto companies, including Binance, still do not have access. Security specialists in the crypto sector argue defenders also need advanced AI models to keep pace with increasingly sophisticated attackers. Recent incidents have included suspected AI-assisted vulnerability attacks targeting Boltz, a non-custodial Bitcoin swap platform, and Coinkite, the maker of the Coldcard crypto wallet, as related damage rises.