406 researchers call for stronger containment of cyber-capable AI models. Yoshua Bengio announced the open letter on his own X account on October 13; the original page also lists Geoffrey Hinton, Stuart Russell, Yejin Choi and Bernhard Schölkopf.
Organizations should initially default to network isolation and adopt high-assurance containment. The proposal concerns cyber-capable models, rather than every AI application indiscriminately.
The signatories propose an international body to standardize practices and track containment failures. Security incidents should be reported to it and eventually made public.
Their expectation of more damaging future incidents remains a warning. The letter does not establish that body. The figure 406 describes the list visible at retrieval, rather than an immutable final count.