AffirmativeAI: Towards LGBTQ+ Friendly Audit Frameworks for Large Language Models

TL;DR

The paper addresses the risk that general-purpose large language models used for LGBTQ+ mental health support may deliver non-affirming or unsafe responses. It introduces an LGBTQ+ Affirmative Framework based on the 3As—Affirmative Attitude, Accurate Knowledge, and Appropriate Action—and proposes prompt-based evaluation to test LLMs against diverse LGBTQ+ scenarios. A therapist-informed benchmarking approach combines qualitative feedback and quantitative metrics to assess affirmativeness, aiming to identify harms and guide safer deployment. By aligning LLM interactions with affirmative therapy principles, the work seeks to enable more supportive, informed, and context-sensitive AI assistance for LGBTQ+ individuals while avoiding overreliance on digital tools for mental health care.

Abstract

LGBTQ+ community face disproportionate mental health challenges, including higher rates of depression, anxiety, and suicidal ideation. Research has shown that LGBTQ+ people have been using large language model-based chatbots, such as ChatGPT, for their mental health needs. Despite the potential for immediate support and anonymity these chatbots offer, concerns regarding their capacity to provide empathetic, accurate, and affirming responses remain. In response to these challenges, we propose a framework for evaluating the affirmativeness of LLMs based on principles of affirmative therapy, emphasizing the need for attitudes, knowledge, and actions that support and validate LGBTQ+ experiences. We propose a combination of qualitative and quantitative analyses, hoping to establish benchmarks for "Affirmative AI," ensuring that LLM-based chatbots can provide safe, supportive, and effective mental health support to LGBTQ+ individuals. We benchmark LLM affirmativeness not as a mental health solution for LGBTQ+ individuals or to claim it resolves their mental health issues, as we highlight the need to consider complex discrimination in the LGBTQ+ community when designing technological aids. Our goal is to evaluate LLMs for LGBTQ+ mental health support since many in the community already use them, aiming to identify potential harms of using general-purpose LLMs in this context.