Cybersecurity stocks declined on Friday, with notable drops in companies such as CrowdStrike (-7%), Palo Alto Networks (-6%), Zscaler (-4.5%), and Okta, SentinelOne, and Fortinet (around -3%).
The downturn appears to be linked to reports that Anthropic’s latest AI model was leaked ahead of its official release. According to Fortune, details about the model — internally named “Claude Mythos” — were discovered in a draft blog post stored in a publicly accessible and unsecured system.
Anthropic attributed the incident to human error in the configuration of its content management system. The leaked material introduced a new class of AI models called “Capybara,” which is expected to surpass the company’s current top-tier model, Claude Opus.
Based on the draft, the new model demonstrates superior performance in coding, academic reasoning, and cybersecurity-related tasks. “Claude Mythos” was described as the most powerful AI model the company has developed to date.
Market analysts raised concerns about the broader implications. They noted that advancements like this could weaken traditional cybersecurity defenses, as methods based on known signatures or vulnerability databases may struggle to keep up with AI-driven attacks.
Additionally, experts expect increased attack complexity, higher defense costs, and potential shifts in how organizations allocate cybersecurity budgets.
Some analysts warned that such powerful AI tools could elevate the capabilities of ordinary hackers to levels comparable to nation-state actors, significantly increasing global cyber risk.
On the other hand, these developments may drive further investment in cybersecurity, particularly in AI-powered defense solutions, as organizations seek to counter increasingly sophisticated threats.
0 Comentários