Anthropic recently introduced its latest AI model, Mythos, which boasts advanced capabilities tailored for defensive cybersecurity tasks. The unveiling of Mythos raised concerns about potential risks to conventional software security, as the AI startup disclosed uncovering numerous significant vulnerabilities across major operating systems and web browsers.
Despite skepticism from some experts questioning the authenticity of Anthropic’s claims regarding the potency of their AI technology, even critics acknowledged that Mythos may signify a significant progression in AI development. The model was launched under a controlled initiative named “Project Glasswing,” with select tech giants like Amazon, Microsoft, Nvidia, and Apple gaining access. Additionally, over 40 other organizations involved in critical software infrastructure were granted access to the preview.
Experts have cautioned that Mythos has the ability to identify and exploit previously unknown vulnerabilities at a pace faster than companies can address them. Its advanced coding and autonomous functions could potentially hasten sophisticated cyberattacks, particularly in sectors such as banking that heavily rely on intricate, interconnected, and often outdated technology systems.
Anthropic emphasized the model’s potential to detect software flaws on a large scale, highlighting the serious risks to economies, public safety, and national security if misused. Following the Mythos launch, U.S. software stocks experienced a decline, reflecting concerns that AI advancements could disrupt traditional firms.
Regulators, including Bank of Canada Governor Tiff Macklem, stressed the need for financial systems globally to address the risks posed by rapid AI advancements like Mythos. Discussions surrounding Mythos have taken place among financial sector resiliency groups, indicating a growing awareness of the implications of such advanced AI technologies. Macklem underscored the necessity for comprehensive plans to manage the evolving landscape of AI technologies and their potential impact on cybersecurity.
The White House engaged in discussions with Anthropic CEO Dario Amodei regarding Mythos, focusing on collaboration, cybersecurity, and balancing AI innovation with safety measures. Despite concerns, the U.S. government is reportedly planning to provide a version of Mythos to major federal agencies. Similarly, British authorities and major banks in Germany have initiated talks to assess the possible risks associated with Mythos.
In light of the concerns raised, influential voices like David Sacks, a notable Anthropic critic and former White House AI and crypto czar, advised taking the situation seriously. Sacks acknowledged the increasing capabilities of coding models like Mythos in identifying vulnerabilities, potentially leading to the exploitation of multiple vulnerabilities and the creation of exploits.