OpenAI announced on November 17 that it is working with the NATO Communications and Information Agency to provide the alliance’s core bodies with access to advanced cybersecurity models through the Daybreak program.
NATO’s monthly usage could reach up to 300 billion input and output tokens. Authorized tasks include identifying software weaknesses, validating vulnerabilities, and developing and testing remediation plans. OpenAI did not disclose the specific models, the start date of the arrangement, or the financial terms; related testing showed that the models can help analyze vulnerabilities, but researchers must still validate hypotheses, rule out incorrect conclusions, and assess the impact.
Source:Dataconomy · Read original
Leave a Reply