Anthropic offers METR researchers in-depth insight

Anthropic, the developer of advanced AI models, has decided to grant METR researchers employee-level access to its systems. This decision allows METR to gain in-depth insight into the practices and risks associated with Anthropic's AI technology.

An important aspect of this collaboration is that METR will publish its own assessments of the risks, incidents, and practices within Anthropic. This not only provides transparency but also strengthens Anthropic's accountability in the use of its technology. The organization has also waived its right to block these publications.

This step can be seen as an effort to ensure safety and accountability within the AI sector. By allowing independent evaluations, Anthropic hopes to enhance the trust of users and stakeholders while simultaneously gaining valuable insights into the impact of their systems.

It is a progressive approach that can serve as an example for other companies in the sector. As AI technologies continue to evolve, it becomes increasingly important to ensure transparency and accountability. This way, both developers and users can better understand how AI works and the associated risks.

Read the full article from Techcrunch.