Follow the latest coverage, related explainers and connected technology stories.
Two officials at the U.S. Department of Defense warned that delays in modernizing the Pentagon’s digital infrastructure have made it more vulnerable to highly destructive cyberattacks as artificial intelligence rapidly evolves. The article is based on a report published by the Washington Post, while no independent technical details were presented about the nature of the vulnerabilities or the attacks in question.
An opinion article discusses Dario Amodei’s warning, the CEO of Anthropic, that swarms of artificial intelligence agents could cause massive economic damage within 6 to 12 months if their capabilities develop without sufficient safeguards. The article presents this warning as a possibility based on the OpenAI-Hugging Face incident, not as a confirmed prediction.
TechCrunch discusses the rising warnings that artificial intelligence could pose an existential threat to humanity, and whether these warnings stem from genuine concerns or are also becoming a way to showcase model capabilities and enhance companies’ value. The discussion raises questions about companies’ control over their models and how this language could affect Anthropic’s potential IPO.
Anthropic said it detected the use of its artificial intelligence models in activities that could support the development of biological weapons, including planning experiments on the adaptation of avian influenza to mammals and computationally redesigning toxins. The company announced that it had banned the accounts involved and strengthened its safeguards, while emphasizing that detecting these cases does not necessarily prove criminal intent on the part of those involved.