Human-AI Collaboration Enables More Empathic Conversations in Text-based Peer-to-Peer Mental Health Support (2203.15144v1)

Abstract: Advances in AI are enabling systems that augment and collaborate with humans to perform simple, mechanistic tasks like scheduling meetings and grammar-checking text. However, such Human-AI collaboration poses challenges for more complex, creative tasks, such as carrying out empathic conversations, due to difficulties of AI systems in understanding complex human emotions and the open-ended nature of these tasks. Here, we focus on peer-to-peer mental health support, a setting in which empathy is critical for success, and examine how AI can collaborate with humans to facilitate peer empathy during textual, online supportive conversations. We develop Hailey, an AI-in-the-loop agent that provides just-in-time feedback to help participants who provide support (peer supporters) respond more empathically to those seeking help (support seekers). We evaluate Hailey in a non-clinical randomized controlled trial with real-world peer supporters on TalkLife (N=300), a large online peer-to-peer support platform. We show that our Human-AI collaboration approach leads to a 19.60% increase in conversational empathy between peers overall. Furthermore, we find a larger 38.88% increase in empathy within the subsample of peer supporters who self-identify as experiencing difficulty providing support. We systematically analyze the Human-AI collaboration patterns and find that peer supporters are able to use the AI feedback both directly and indirectly without becoming overly reliant on AI while reporting improved self-efficacy post-feedback. Our findings demonstrate the potential of feedback-driven, AI-in-the-loop writing systems to empower humans in open-ended, social, creative tasks such as empathic conversations.

Overview of Human-AI Collaboration in Enhancing Empathic Conversations

The paper "Human-AI Collaboration Enables More Empathic Conversations in Text-based Peer-to-Peer Mental Health Support" outlines an investigation into augmenting peer-to-peer mental health support with AI-driven feedback mechanisms. The authors explore challenges associated with facilitating empathic interactions in online mental health support forums and present Hailey, an AI-in-the-loop system designed to enhance users' empathic communication.

Key Findings and Methodologies

The authors describe a randomized controlled trial involving 300 participants from TalkLife, an online peer-to-peer mental health support platform. The deployment of Hailey resulted in a notable 19.60% improvement in expressed empathy in peer supporter responses and even more significant gains (38.88%) within the subgroup that found providing support challenging. Hailey offers just-in-time suggestions to improve empathy in textual responses. This approach focuses not on generating responses from scratch, but instead on refining human-generated content—a strategy aimed at maintaining the authenticity of human interactions.

Implications for Human-AI Collaboration

Empathy is a critical component of mental health support, and the results underscore the potential of AI-assisted technologies in enhancing communication quality in non-clinical environments. The system empowers users to engage in more meaningful interactions, which is especially valuable given the understaffing issues prevalent in mental healthcare. Such AI systems could offer feasible alternatives to traditionally labor-intensive empathy training, making support more accessible and scalable.

Future Directions and Challenges

A significant area for future development highlighted by the authors is ensuring such AI systems do not unintentionally infringe upon the emotional authenticity of human interactions. The balance between AI influence and human autonomy remains a pivotal challenge, with emphasis needed on developing AI systems that support, rather than redefine, human empathic capacities. Moreover, the paper points to the potential secondary benefits for peer supporters, including improved confidence and skill development in providing support.

Conclusion

This research contributes to ongoing dialogues within AI and mental health support, demonstrating the viability of feedback-driven Human-AI collaboration to enhance empathic discourse in online support settings. It opens avenues for broader application of AI systems in high-risk tasks while also addressing ethical considerations necessary for such integration. Continued refinement and testing, particularly across diverse socio-cultural contexts, will be essential for optimizing and scaling AI interventions like Hailey.