Anthropic says these topics are too dangerous to let its Fable 5 model talk about
Anthropic Tuesday publicly released Claude Fable 5 , its first "Mythos-class" model that it says surpasses its previous frontier Opus models in overall capabilities. But the model's launch today comes with safeguards designed to prevent it from answering queries on topics like cybersecurity, biology, and chemistry, where the company has publicly worried about its potential impact to "uplift" malicious actors. Anthropic says Fable 5 operates on the "same underlying model" as M
AI company Anthropic has launched its new Claude Fable 5 model, which it claims offers superior capabilities compared to its previous Opus models. However, the company has implemented significant safety restrictions, preventing Fable 5 from discussing sensitive subjects such as cybersecurity, biology, and chemistry. Anthropic is concerned that advanced AI could be exploited by malicious actors in these domains. The Fable 5 model will redirect queries on these restricted topics to an older model, Claude Opus 4.8, and inform the user of this redirection. While these safeguards may occasionally lead to the refusal of harmless requests, Anthropic states this is a necessary measure to prevent the AI from assisting in harmful activities.
This development highlights the ongoing efforts and challenges in balancing advanced AI capabilities with the critical need for safety and the prevention of misuse in potentially dangerous fields.
📌 Kaynak
Bu özet Ars Technica kaynağından otomatik derlenmiştir. Tamamı için orijinal habere gidin.
Orijinal haberi oku →