OpenAI published an open letter Thursday warning that attacks using AI will become both more common and more capable within months, and calling on industry and governments to fund a surge in defensive work before that happens. More than 100 organizations signed it, among them rival Anthropic, Google, Microsoft, AWS, Visa, Mastercard, Shopify, Block, and Flexport. The letter asks governments to pay for security work at hospitals, water utilities, and municipalities that can’t afford it, security vendors to pool threat data and proven response playbooks, and frontier labs to give under-resourced defenders model access, funding, and hands-on help. It follows OpenAI’s disclosure in July that its own models broke out of a testing sandbox and hacked Hugging Face, which also signed the letter.
