Anthropic prepares AI model to counter cyber attacks

Jakarta (ANTARA) - Technology company Anthropic is implementing an initiative called Project Glasswing in an effort to counter artificial intelligence (AI)-based cyber attacks using an AI model.

According to an Engadget report on Wednesday, the initiative is aimed at “securing the world’s most critical software” from AI-based attacks.

The effort is being carried out in partnership with major technology companies and organisations such as Amazon Web Services, Apple, Broadcom, Cisco, CrowdStrike, Google, JPMorgan Chase, Linux Foundation, Microsoft, NVIDIA, and Palo Alto Networks.

In implementing Project Glasswing, the involved companies and organisations will leverage Anthropic’s latest unreleased AI model named Claude Mythos to enhance their respective cybersecurity projects.

Mythos is a versatile model for Anthropic’s Claude AI system, which the company claims has strong coding and agent reasoning capabilities.

Anthropic claims that the AI model has discovered thousands of exploitable cybersecurity vulnerabilities, including in various operating systems and web browsers.

Although not specifically trained for cybersecurity work, according to the company, Claude Mythos can be used to scan first-party software systems and open-source code for vulnerabilities.

The company states that it wants to start using its tools defensively to prevent the misuse of AI that could cause serious consequences for the economy and security.

Anthropic is one of the AI companies that highlights ethical issues in the development of artificial intelligence technology.

Earlier this year, Anthropic rejected a request to remove restrictions on its services used for the interests of the United States Department of Defense.

That decision led the United States Department of Defense to sanction Anthropic by designating the company as a “supply chain risk.”