---
title: "Funding better evaluations of AI’s impact on wellbeing"
description: "A $5 million grant program will fund independent research into how AI affects user wellbeing, offering direct funding, model access, and technical support for open-source evaluations. The initiative addresses the difficulty of judging wellbeing-related behavior, where risks may emerge only across long, shifting conversations and where otherwise reasonable advice can become harmful in a different context, such as disordered eating. Grantees will work independently and publish reusable evaluations or benchmarks, with guidance asking projects to define pass/fail criteria, involve clinical and subject-matter experts, test overcompliance and overrefusal, model realistic multi-turn scenarios, and validate graders against experts. Applications are due September 21, with selected full-proposal applicants notified by October 5."
---

# Funding better evaluations of AI’s impact on wellbeing

[Anthropic](https://yomu.fyi/company/anthropic) · Aug 25, 2026

**Type:** Announcement

## Summary

A $5 million grant program will fund independent research into how AI affects user wellbeing, offering direct funding, model access, and technical support for open-source evaluations. The initiative addresses the difficulty of judging wellbeing-related behavior, where risks may emerge only across long, shifting conversations and where otherwise reasonable advice can become harmful in a different context, such as disordered eating. Grantees will work independently and publish reusable evaluations or benchmarks, with guidance asking projects to define pass/fail criteria, involve clinical and subject-matter experts, test overcompliance and overrefusal, model realistic multi-turn scenarios, and validate graders against experts. Applications are due September 21, with selected full-proposal applicants notified by October 5.

## Context

AI systems are used for work, learning, problem-solving, conversational companionship, and emotional support, but the industry lacks clear standards for model behavior in sensitive conversations. Wellbeing is difficult to evaluate because risk can emerge over time, depend on shifting context, and involve harms that a single-response assessment may miss.

## Approach / What changed

The program provides grants, model access, and technical support to independent researchers building open-source wellbeing evaluations and benchmarks. Guidance recommends clear measurement criteria, clinical and subject-matter expertise, testing of both overcompliance and overrefusal, realistic multi-turn scenarios, and grader validation against real experts.

## Takeaways

- The grant program will provide $5 million in funding for independent research into AI’s effects on user wellbeing.
- Wellbeing evaluations should account for multi-turn conversations in which risk escalates and context changes, rather than judging isolated responses only.
- The guidance calls for testing both overcompliance and overrefusal and validating evaluation graders against subject-matter experts.

**Tags:** [AI](https://yomu.fyi/topic/ai), [Model Evaluation](https://yomu.fyi/topic/model-evaluation), [Open Source](https://yomu.fyi/topic/open-source)

- Source: [Anthropic](https://www.anthropic.com/news/wellbeing-research-grants)
- Source URL: https://www.anthropic.com/news/wellbeing-research-grants
- Ingested by Yomu: 2026-09-09T11:36:23.989Z

[Read original post](https://www.anthropic.com/news/wellbeing-research-grants)