Anthropic restricts advanced AI Claude Fable's biology answers
AI firm Anthropic has intentionally restricted its new Claude Fable model from answering basic biology questions, redirecting them to an older model. Despite being marketed as powerful, Anthropic cites safety concerns related to the model's Mythos-class capabilities, deeming it too dangerous for unrestricted public release, even for seemingly harmless queries.
Key points
- Anthropic released its new AI model, Claude Fable, but intentionally restricted its ability to answer basic biology questions.
- The model redirects such queries to an older version, Claude Opus 4.8, despite possessing the capability to answer them.
- Anthropic classifies Fable as a Mythos-class model, deeming its cybersecurity capabilities too dangerous for unrestricted public release.
- The company cited safety concerns as the reason for limiting even seemingly harmless biological and medical queries.
- This restriction aims to allow customers to benefit from the model's capabilities sooner while mitigating potential risks.
Artificial intelligence company Anthropic has implemented specific limitations on its newly released Claude Fable model, preventing it from responding to fundamental biology inquiries. Users attempting to ask questions typically answered by high school students are instead being redirected to the previous flagship model, Claude Opus 4.8.
Anthropic explained that these restrictions are intentional, despite Fable's advanced capabilities. The model belongs to Anthropic's Mythos-class family, which the company has previously identified as potentially too dangerous for widespread public access due to advanced cybersecurity applications. While cybersecurity risks have been a focus, the most apparent limitations are in Fable's handling of biological information.
Even questions considered far removed from any plausible safety risk, such as defining cell membranes, mitochondria, prions, or the mechanism of mRNA vaccines, were refused by Fable. Similar restrictions were applied to ordinary medical queries, including explanations for hay fever or how asthma medication functions. Anthropic stated this tradeoff was made to allow earlier customer access to the model's strengths while managing associated risks.
Sources
The WireByte editorial team synthesises technology news from multiple primary sources, verifies the facts, and links every source. Articles are produced with AI assistance and reviewed under our editorial policy.