Anthropic launches Project Glasswing, an effort to prevent AI cyberattacks with AI

We spot a batch of punishment and gloom astir nan imaginable antagonistic impacts of artificial intelligence, peculiarly centered connected really it could create caller problems successful cybersecurity. Anthropic has announced a caller inaugural called Project Glasswing to thief reside those concerns by moving "to unafraid nan world’s astir captious software" against AI-powered attacks. The endeavor includes Amazon Web Services, Apple, Broadcom, Cisco, CrowdStrike, Google, JPMorganChase, nan Linux Foundation, Microsoft, NVIDIA and Palo Alto Networks arsenic partners.

Participants will usage Claude Mythos Preview, an unreleased, general-purpose exemplary from Anthropic, to heighten their ain information projects. Anthropic claims that this exemplary has recovered thousands of exploitable vulnerabilities, "including immoderate successful each awesome operating strategy and web browser." The institution said it wants to statesman utilizing its devices defensively to forestall malicious usage of AI that could origin terrible consequences for economies and security.

Anthropic has go 1 of nan notable AI companies raising concerns astir morals successful nan field. Earlier this year, nan business refused to region guardrails connected its services for usage by nan Pentagon, which prompted nan Department of Defense to punishment Anthropic pinch a "supply concatenation risk" nickname successful retaliation. Launching Project Glasswing could beryllium a adjuvant commencement toward improved cybersecurity successful nan AI era, but immoderate harm has already been done. Its ain Claude was reportedly utilized by a hacker against multiple authorities agencies successful Mexico successful February.