Standards consortium · Research orgs · Est. 2025
AI Evaluator Forum (AEF) is a standards consortium of independent AI research and evaluation organizations focused on improving the credibility of third-party evaluations of advanced AI systems. It positions independent evaluation as a public-interest mechanism to build trust in claims about model capabilities and risks, and it works in two closely related directions: (1) establishing voluntary best practices/standards for how independent evaluations should be conducted (including evaluator independence, access, and transparency), and (2) fostering coordination among evaluator organizations so that shared methodologies and transparency norms strengthen the broader evaluation ecosystem over time.
AEF’s launch communications emphasized that AI governance increasingly relies on third-party evaluations, but that quality and impartiality are not guaranteed merely by “third-party” status—AEF’s standards aim to make the operating conditions of evaluations more legible to external stakeholders.
No indexed openings right now.
Publications are not yet indexed for this organization. Read them on aievaluatorforum.org ↗.
AEF announced its creation and described its mission to coordinate independent AI evaluators, while also releasing its initial materials, including a public statement and AEF-1.
Transparency about Third-Party AI Evaluations (public letter initiative page)AEF published an initiative page describing expectations for third-party evaluators to disclose evaluation methods/control, editorial control, access/time, and conflicts of interest to improve external scrutiny and interpretation of evaluation results.
Minimum Operating Conditions for Independent Third Party AI Evaluations (AEF-1 initiative page)AEF described the purpose of AEF-1 and how evaluators can use it as a voluntary standard to document independence, access, and transparency in third-party AI evaluations.