Anthropic, which claimed AI model was too risky for public to use, releases ‘safe’ version

Fonte: New York Post

After hyping potential safety threats posed by its latest AI model, Anthropic says it is releasing a version with tweaks that make it “safe” for the general public. A day after filing for its initial public offering, the company on Tuesday trumpeted a new version of the Mythos model with guardrails around cybersecurity and biological weapons. If the model, called Claude Fable 5, is asked about off-limits issues like exploiting a software bug or creating bioweapons, the response will be blocked and users will be directed to an older model known as Opus 4.