On February 27, 2025, Anthropic launched a Transparency Hub intended to provide a detailed overview of concrete measures the company says it is taking to keep its systems safe, beneficial, and trustworthy.
Anthropic framed the Hub as a response to the rapid advance of AI capabilities and the need for meaningful transparency to support trust and accountability. The company emphasized that organizations working on frontier AI have responsibilities to disclose information about responsible scaling, safety protocols, and risk mitigation, particularly as regulatory frameworks evolve.
Initial periodic report: key metrics
As part of the Transparency Hub, Anthropic published its first periodic report covering a set of transparency metrics. The report includes disclosure of the following metrics:
- number of banned accounts
- account appeals
- appeal overturns
- reports to the National Center for Missing and Exploited Children (NCMEC)
- government request data
The announcement does not provide specific numeric values for these metrics; the emphasis is on initiating regular reporting of these categories.
What the Hub contains
The Transparency Hub offers detailed information on several topics:
- methodologies for model evaluation and safety testing
- platform abuse detection and enforcement measures
- internal governance and risk assessment policies
- methods for assessing and addressing potential societal impacts
- research initiatives advancing AI safety
- security and privacy safeguards applied during development
Unified framework and ongoing commitment
Anthropic described the Hub as a unified framework designed to address the fragmentation of documentation requirements across multiple transparency initiatives and voluntary commitments. The company said the Hub aims to give users, policymakers, and other stakeholders a structured, accessible, and accountable view into model development and deployment.
Anthropic also committed to expanding its reporting over time to reflect evolving best practices and emerging challenges as AI capabilities develop. The company invited feedback at transparency@anthropic.com.
