'Project Glasswing' has allowed 150 businesses in over 15 countries, including India, access to Anthropic's cybersecurity AI model Mythos. Organisations that take part in the programme are paired with the closed Claude Mythos Preview model. Pairing helps find important zero-day vulnerabilities early on, before attackers have a chance to exploit them.
There were concerns about potential exploitation following the model's April release, which led to strong restrictions on the model's use. The model is intended to find software security issues, and early testing revealed that it might quickly disclose thousands of vulnerabilities.
Mythos Expanding its Reach
About fifty partner organisations were initially part of the Glasswing Project. Among these groups are prominent tech corporations like Amazon, Google, Microsoft, Apple, and NVIDIA; cybersecurity businesses like CrowdStrike and Palo Alto Networks; and the AI Security Institute of the United Kingdom, which was involved in the early access program. According to Anthropic, these early adopters had found more than 10,000 critical security flaws in Mythos by the end of May.
As a result of the most recent extension, organisations from South Korea, Japan, Australia, Canada, France, Germany, Italy, Switzerland, the Netherlands, Spain, Belgium, and Spain will all have access to Mythos. With this move, Anthropic is trying to sell Mythos to governments, vital infrastructure operators, and important organisations as a cybersecurity tool. However, there are growing concerns about its potential risks and its ability to uncover software flaws around the world.
|
Some Interesting Facts of the Story |
|
1.The model reportedly demonstrated the 2.The project's name, 3.The initiative showcases how AI is |
India Shaping Strong AI Shield to Thwart Cybersecurity Threats
India increased its efforts to detect software vulnerabilities in response to recent cybersecurity concerns. India reportedly tested critical government and financial software against the Mythos AI model a few days ago. Reports surfaced that prominent IT companies such as Infosys and Tata Consultancy Services were allegedly performing controlled security testing.
Critical systems like Aadhaar and government login platforms were, however, being reviewed by CERT-In. These organisations were utilising Anthropic's Claude Opus 4.7 to find and repair vulnerabilities since they couldn't access Mythos. Amid jurisdictional, compliance, and national security concerns regarding foreign-hosted infrastructure utilised in sensitive sectors like banking, telecom, and critical infrastructure, the Indian central government is allegedly advocating for the sovereign hosting of the Claude AI model in India. In addition, representatives from the Ministry of Finance, MeitY, and CERT-In met with Anthropic to discuss the Mythos model.
|
Quick Shots |
|
•Anthropic has expanded Project Glasswing •The initiative provides selected •Mythos is aimed at identifying critical •Early users of the model reportedly |
Original Article
(Disclaimer – This post is auto-fetched from publicly available RSS feeds. Original source: Startuptalky. All rights belong to the respective publisher.)
