Claude Mythos Preview initially refused to work on the AES cryptanalysis task, stating that further improvements were impossible, before being prompted by researchers to seek novel ideas

From the articleAI Finds New Weaknesses in Cryptographic Algorithms, Anthropic SaysAlso in Anthropic AI model finds flaws in tough-to-crack encryption algorithms, Claude AI Just Cracked a Post-Quantum Test Scheme and Found a Faster 7-Round AES Attack, Anthropic's Claude Mythos ‘breaks’ AES encryption that protects the entire internet; including online bank transactions, online chats and more, Anthropic says its Mythos model found vulnerabilities in cryptographic algorithms that secure the internet, Claude Mythos Cracked Post-Quantum Cryptography That Humans Spent Years Failing to Break - Decrypt