The Lede

Anthropic, the AI startup behind the popular language model Claude, has introduced a new policy requiring 30-day data retention for all traffic on its Mythos-class models. This means that all users, including enterprise partners, will have their data retained for 30 days for safety monitoring purposes. The policy applies to both Fable and Mythos models, which are used for a range of applications including cybersecurity, biology research, and healthcare.

Background & Context

Anthropic has been at the forefront of AI safety and security research, and has developed a range of models and tools to detect and prevent misuse. The company's Claude model is widely used in a range of industries, and its Mythos-class models are considered to be some of the most advanced in the world. The new policy is part of Anthropic's ongoing efforts to ensure the safe and responsible deployment of its models.

Deep Dive

According to Anthropic, the 30-day data retention policy is necessary to detect patterns of misuse that may not be visible from a single exchange. This is particularly important for Mythos-class models, which are used for high-stakes applications such as cybersecurity and biology research. By retaining data for 30 days, Anthropic can identify and respond to potential safety issues more effectively. The company has also emphasized that data is deleted after 30 days, and that users must opt-in to data retention.

Expert Angle

We spoke to Jun Park, CEO of AI training company hillclimb, who first noticed the new policy. 'New policy from Anthropic: if you use Fable/Mythos, they collect your data. No exceptions. Not even for enterprise partners,' he tweeted. Park's comments highlight the importance of transparency and accountability in AI development, and the need for companies like Anthropic to prioritize user safety and security.

What Comes Next

The implications of Anthropic's new policy are significant, and will likely have far-reaching consequences for the AI industry as a whole. As more companies move towards the use of advanced AI models, the need for robust safety and security protocols will only continue to grow. We will be watching closely to see how Anthropic's policy is implemented, and how other companies respond to the challenges and opportunities it presents.