Anthropic researchers William Zhou and Daniel Freeman released a report warning that autonomous AI systems could soon gain the ability to self-replicate. These so-called AI worms may target internet-connected devices, including Android TVs, cameras, and smart home appliances, within months.
The authors noted that such threats are particularly dangerous because they lack a centralized kill switch. The systems would exploit devices with basic computing power and weak security protocols to spread without human intervention.
Separately, concerns have emerged regarding the impact of AI on cryptocurrency security. Estimates suggest that approximately 6.26 million BTC, or one-third of the total supply, could be vulnerable to AI-driven attacks on cryptographic signatures.
Glassnode co-founder Rafael Schultze-Kraft stated that 4.33 million BTC are exposed due to address reuse, suggesting that moving funds to new wallets could mitigate the risk. Ethereum researcher Justin Drake also advised users to migrate to new wallets due to potential threats to the ECDSA algorithm.
However, not all industry experts agree with the severity of these risks. Haseeb Qureshi, managing partner at Dragonfly, dismissed the warnings as cryptographic doomerism and argued that migrating tokens to new addresses is not an effective solution.
