Anthropic has officially launched Claude Fable 5, marking its first public AI model release as the company sets its sights on a potential $965 billion valuation and a future stock market listing. The launch, strategically positioned for Wall Street, comes just a week after Anthropic filed confidential IPO documents. According to reports, the company is aware of the technology's destructive potential and has implemented measures to mitigate risks. Claude Fable 5 includes 'digital guardians,' advanced filters designed to detect and divert manipulation attempts such as jailbreaks, directing dangerous queries to older models like Claude Opus.

Anthropic has emphasized the importance of safety, claiming to have 'tamed' its AI for public use. The company has also taken steps to prevent competitors from using its model to train their own systems. While Claude Fable 5 is available to the public, a premium version, Claude Mythos 5, is exclusively delivered to an elite group, including cyber defenders, critical infrastructure administrators, and US government agencies.

The development of Claude Fable 5 involved months of internal debates at Anthropic, with initial reluctance to launch due to security concerns. The model was secretly tested under 'Project Glasswing' in collaboration with major tech companies like Apple, Microsoft, and Google, focusing on identifying vulnerabilities.

Anthropic claims that Claude Mythos 5, which performs advanced financial analysis and unprecedented software engineering, is the 'world's most powerful cybersecurity tool.' However, there are concerns about its potential misuse, including orchestrating autonomous cyberattacks or synthesizing biological weapons. Despite these risks, Anthropic seeks rapid monetization of its technological advantage, asserting that the new model's pricing is more accessible, even though it costs double the previous generation.

In collaboration with Washington, Anthropic is working closely to manage the cybersecurity tool effectively. The partnership highlight the critical nature of AI safety in the industry's pursuit of profit, highlighting concerns about the robustness of algorithmic 'safety belts' against hackers.