'Is Anthropic the new Dr Frankenstein?' | BBC News - Modern Ghana

Anthropic's AI model Claude Mythos has sparked controversy after it was deemed too dangerous to release, according to a BBC report. The report, on the programme AI Decoded, features marketing professor Scott Galloway, who compared Anthropic to Dr Frankenstein. The model uncovered "thousands of hidden vulnerabilities" and can outperform humans at certain hacking tasks, raising concerns about its potential misuse.

The decision not to release Claude Mythos was made by Anthropic itself, citing safety concerns. The model's capabilities include identifying security flaws at a scale and speed beyond human capacity. The BBC's AI Decoded programme explored the ethical implications of building such powerful systems without clear deployment plans.

Anthropic has positioned itself as a safety-focused company, and the Claude Mythos case highlights the challenges of balancing capability and caution. The controversy comes as AI companies increasingly develop models with advanced hacking skills, prompting debates about whether such systems should be kept private or open-sourced.