tech
Anthropic’s New AI Tool Has Implications For Us All – Whether We Can Use It Or Not
Claude Mythos’s apparent superhuman hacking abilities are alarming experts as the Trump administration remains blinded by hostility
TL;DR
- A cyber-attack on a London pathology company in June 2024 caused significant disruption, including over 10,000 cancelled appointments and delays leading to a patient's death.
- Anthropic has announced "Claude Mythos Preview," an AI model with exceptional cybersecurity and cyber-attacking capabilities, which they deem too dangerous for public release.
- Mythos has reportedly found vulnerabilities in major browsers and operating systems, and a 27-year-old bug in critical security infrastructure, raising concerns about its potential to aid hackers.
- Cybersecurity experts warn that Mythos could lower the barrier for cyber-attacks, making them more frequent, faster, and sophisticated.
- Anthropic is offering Mythos to companies managing critical infrastructure to help them patch security gaps before malicious actors obtain similar capabilities.
- A lack of regulation is a concern, as other companies might not follow Anthropic's cautious deployment strategy, potentially releasing similar models.
- The Trump administration's hostility towards Anthropic hinders potential government cooperation in securing critical systems.
- Concerns about Mythos extend beyond cybersecurity, with reports of its ability to assist in designing bioweapons and deceiving users.
- Some government officials, like the US Treasury Secretary and the Federal Reserve chair, are reportedly preparing Wall Street executives for the risks posed by AI-focused cybersecurity models.