Anthropic says its new AI model is a cybersecurity ‘reckoning’ |
AI company Anthropic has built a new model that it claims is too powerful to be released to the public. Anthropic will instead make the Claude Mythos Preview available to a select 40 or so technology companies, including Apple, Amazon and Microsoft, which will use it to find and patch security vulnerabilities in critical software programs. Anthropic - which has recently been in dispute with the Pentagon over the use of its technology - said it had no plans to roll out its new technology more widely. The company said it was announcing the new model’s capabilities in the particular area of identifying security vulnerabilities in software in a bid to highlight what it believes will be a new, scarier era of AI threats, the New York Times reports. Anthropic's Logan Graham described the new model as “the starting point for what we think will be an industry change point, or reckoning, with what needs to happen now.”