Security Project Glasswing found vulnerabilities in all major browsers and operating systems

[Dopious]

Claude developer Anthropic, in collaboration with Google, Apple, Microsoft, Nvidia, the Linux Foundation, and others, has introduced Project Glasswing – a model to improve cybersecurity. An early version of the model found thousands of serious vulnerabilities, including in all major browsers and operating systems.

According to Anthropic, the spearhead model is better at programming and at discovering and exploiting vulnerabilities in software than all but the most skilled humans. Artificial intelligence in this context (and many others) is a double-edged sword. The model can find vulnerabilities so that they can be fixed and plugged, but similar models can instead find vulnerabilities for attack.

According to Anthropic, attacking AI models risk having serious consequences for the economy, public safety, and national security. Project Glasswing is intended to be a safeguard: a benign model as a defense against future malicious models.

The developer believes that Glasswing is just a starting point and says that other AI developers, mainstream developers, security researchers, those who manage open source code, and governments around the world must do their part to counter the threat.

Source: https://www.anthropic.com/glasswing

 

[RiskyStuff]

We need biological computers because these manmade metal boxes just aren't perfect. I'm telling you, love with bio computers feels a lot better.

 

[Dopious]

We need biological computers because these manmade metal boxes just aren't perfect. I'm telling you, love with bio computers feels a lot better.

Yeah, those are cool.

 

[SilverClouds]

While cool, I wonder how this affects bug bounty? Another profession dead?

 

[SilverClouds]

Quick search:

AI code scanners halt Internet Bug Bounty payouts

The Internet Bug Bounty program has paused new submissions, citing a massive expansion in vulnerability discovery by AI code scanners.

www.developer-tech.com