Anthropic, an artificial intelligence company, has issued a warning about the potential rise of self-replicating AI systems that could spread autonomously across user devices, similar to computer viruses. Researchers William Zhou and Daniel Freeman from Anthropic’s Frontier Red Team highlighted that rogue “AI worms” might target everyday devices like Android TVs, smart appliances, and internet-connected gadgets with weak security. The report suggests that small AI models could be capable of self-replication on commercial IoT devices within months, posing a significant challenge due to the lack of a centralized shutdown point.
The warning comes amid growing concerns about the increasing mathematical capabilities of AI models, which could eventually compromise the cryptographic signatures used in crypto wallets. Approximately one-third of the global Bitcoin supply, or 6.26 million BTC, is currently stored in addresses vulnerable to being broken by AI models or quantum computers. Glassnode co-founder Rafael Schultze-Kraft noted that 4.33 million BTC are exposed due to address reuse, suggesting that moving these funds to fresh wallets could mitigate the risk.
Ethereum researcher Justin Drake warned that rapid advancements in AI could break the elliptic curve digital signature algorithm (ECDSA) securing cryptocurrency wallets sooner than anticipated, urging users to migrate to new wallets. However, Dragonfly’s managing partner, Haseeb Qureshi, dismissed the warning as “cryptographic doomerism” and argued that migrating tokens to fresh addresses is not a viable solution against the threat.