We use cookies

We use cookies to ensure you get the best experience on our website. For more information on how we use cookies, please see our cookie policy.

Tech News
Back to home

Articles tagged with "AI, cybersecurity, vulnerabilities"

Anthropic's Claude Mythos AI has discovered thousands of vulnerabilities in every OS and browser
News

Anthropic's Claude Mythos AI has discovered thousands of vulnerabilities in every OS and browser

Anthropic's Claude Mythos AI, a powerful unreleased model, has identified thousands of high-severity vulnerabilities in major operating systems and browsers, surpassing human capabilities in finding and exploiting these flaws. The AI poses a significant cybersecurity threat, prompting concerns about potential misuse by malicious actors. While Claude Mythos won't be publicly released, it is being used in the Project Glasswing initiative to secure critical software, with partners like Amazon Web Services, Microsoft, and Google utilizing its capabilities. Anthropic plans to share its findings with the security industry, emphasizing the importance of responsible AI deployment and potential regulation to mitigate risks.

TweakTown•4/8/2026
00
Anthropic's latest AI model identifies 'thousands of zero-day vulnerabilities' in 'every major operating system and every major web browser' — Claude Mythos Preview sparks race to fix critical bugs, some unpatched for decades
News

Anthropic's latest AI model identifies 'thousands of zero-day vulnerabilities' in 'every major operating system and every major web browser' — Claude Mythos Preview sparks race to fix critical bugs, some unpatched for decades

Anthropic's latest AI model, Claude Mythos Preview, has uncovered "thousands of zero-day vulnerabilities" in major operating systems and web browsers, sparking a rush to fix these critical bugs, some of which have remained unpatched for decades. The lab is collaborating with industry giants and government entities under "Project Glasswing" to proactively address these vulnerabilities before they can be exploited. Mythos' ability to identify and exploit vulnerabilities poses significant security risks, prompting responsible disclosure practices from Anthropic. While the model won't be widely available, the lab aims to prepare partners for a future where such advanced AI models become more common.

Tom's Hardware•4/7/2026
00
Why AI Keeps Falling for Prompt Injection Attacks
News

Why AI Keeps Falling for Prompt Injection Attacks

The article discusses why AI systems are vulnerable to prompt injection attacks, drawing parallels to drive-through ordering systems. Authors Bruce Schneier and Barath Raghavan emphasize the risks associated with large language models (LLMs) and the importance of AI safety in cybersecurity. They highlight the challenges of securing AI systems against agentic AI attacks. The piece underscores the need for improved AI security measures to mitigate prompt injection vulnerabilities effectively.

IEEE Spectrum•1/21/2026
00

No more articles to load