Claude

Not verified. Claim this page

Visit Website

Claude Sustainability Actions

Published a Responsible Scaling Policy

Anthropic says it first published its Responsible Scaling Policy in September 2023. This policy sets out how it plans to add stronger safety steps as its AI models become more capable. The idea is to match the level of protection to the level of risk. The company says the policy is meant to guide internal decisions and also help shape wider industry safety practice. In plain terms, it is a rulebook for when extra checks, security, and deployment limits should be used.

9 16

Updated its safety framework in 2026

Anthropic says it updated its Responsible Scaling Policy in February 2026 to reflect its current view of AI risk and how to reduce catastrophic harm. It also published a Frontier Compliance Framework in December 2025. That framework explains how it assesses risks such as cyber abuse, chemical, biological, radiological, and nuclear threats, plus sabotage, loss of control, and harmful manipulation. This shows the company has formal processes for reviewing high-risk model behavior and for deciding what safeguards to use.

9 16

Uses testing and outside review

Anthropic says it uses regular evaluations, threat modeling, red team testing (deliberate attempts to find weaknesses), and expert consultations to check model risks. It also says it has worked with outside groups such as the UK AI Security Institute, the US Center for AI Standards and Innovation, and METR. The company says it has a Responsible Disclosure Policy, bug bounty programs, and reporting channels for safety issues. These steps are meant to catch problems before and after release and to let outsiders report concerns.

9 16

Earned ISO/IEC 42001 certification

Anthropic says it achieved ISO/IEC 42001:2023 for its AI management system. ISO/IEC 42001 is an international standard for how organisations manage AI responsibly. It focuses on governance, risk control, and ongoing improvement. In practice, this means the company has had its AI management approach checked against a formal outside standard rather than only describing its own processes. This is one of the clearest third-party signals in the source set.

9 16

Positions Claude for ESG work

Several third-party guides describe Claude as useful for ESG reporting, climate risk analysis, carbon accounting, and sustainability report review. These guides say it can help extract data, map disclosures to frameworks like GRI, SASB, TCFD, and CSRD, and draft first-pass reports. This is not a sustainability action by itself, but it shows the product is being used in ESG workflows. The sources also warn that AI outputs still need human checking because the results depend on the quality of the input data.

9 12 13 16

About Claude

  • Status
  • Unverified