BN
|
TechAI Desk•1 views

Anthropic's Mythos AI: Firms Get Defense Tool as Attacks Speed Up

Anthropic is selectively sharing its Mythos AI model with leading tech and cybersecurity companies to enhance cyber defenses. The model, which can identify vulnerabilities faster than humans, is not publicly available due to abuse concerns. Anthropic has briefed U.S. officials on its capabilities and offered support for government testing. Experts note that AI significantly accelerates both attack and defense, but the gap may widen if defenders do not adopt such tools. The company aims to preempt major security flaws in widely used software. This move highlights the accelerating AI arms race in cybersecurity.

Ad slot
Anthropic's Mythos AI: Firms Get Defense Tool as Attacks Speed Up

Anthropic is providing its Mythos AI model to a select group of major technology and cybersecurity firms to bolster defenses against rapidly evolving AI-driven cyber threats, while opting not to release it publicly due to concerns about potential misuse.

Companies Granted Access

The following organizations will have access to Mythos for cyber defense purposes:

  • Amazon
  • Apple
  • Cisco
  • Google
  • JPMorgan Chase
  • Microsoft
  • Broadcom
  • Nvidia
  • Linux Foundation
  • CrowdStrike
  • Palo Alto Networks

This access includes using the model to identify software bugs and test the resilience of products against specific hacking techniques.

Limited Release and Government Engagement

Anthropic has decided against a general public launch for Mythos, citing risks that cybercriminals and spies could abuse its capabilities. The company has briefed senior U.S. government officials on the model's full offensive and defensive cyber capabilities and has offered to support government testing and evaluation efforts.

Capabilities and Industry Implications

Mythos has demonstrated the ability to discover thousands of previously unknown software vulnerabilities at a rate far exceeding human researchers. Experts warn that AI agents can scan for and exploit vulnerabilities much faster and more persistently than human hackers, potentially widening the gap between attackers and defenders. Cybersecurity professionals emphasize that defenders must adopt similar AI tools to maintain parity in the evolving cyber arms race.

Ad slot
Ad slot