Anthropic Set to Launch Mythos AI With New Name & Stronger Safety Controls

Outlook Business Desk

Mythos AI Release

Anthropic, the AI company behind the Claude family of models is reportedly planning to release its advanced Mythos AI model to the public, according to reports. The company may launch it under a new name and add stronger safety measures before wider availability.

Rebranded Launch

The upcoming version of Mythos is expected to be rebranded for public users. While the core system will remain similar, Anthropic plans to apply stricter usage controls to reduce risks linked to misuse and highly advanced outputs.

Safety Concerns

The company did not release Mythos to the public at first, as per reports, because it feared the model could be misused for cyberattacks. Its strong capability to analyse software flaws led to internal concerns about giving open access.

Controlled Testing Program

Instead of a public launch, Anthropic tested Mythos through Project Glasswing. Selected organisations got secure access to assess how the model identifies cybersecurity vulnerabilities in real-world conditions.

AI Generated Image

Global Testing Expansion

Project Glasswing later expanded to more than 15 countries and included over 100 organisations. Companies from India also took part, along with major firms like Apple, Google and Microsoft, for structured testing and evaluation of the model.

Dragos Condrea

Vulnerabilities Found

Anthropic said Mythos identified thousands of serious vulnerabilities in widely used software. These issues impacted operating systems, browsers and critical applications, with some flaws remaining undetected for many years.

AI Generated Image

Cyber Risk Warning

The company also warned that AI systems like Mythos could be used to exploit known but unpatched vulnerabilities. It said attackers may quickly turn these flaws into working exploits using advanced AI tools

What Is Anthropic’s Claude Mythos & Why Is It Raising Cyber Risk Concerns? Read Here

Read More