It Takes Two to Tango: Towards Theory of AI's Mind

TL;DR

This paper emphasizes the importance of humans developing a theory of AI's mind for effective collaboration, demonstrating that minimal training can improve prediction of AI responses in Visual Question Answering, despite current interpretability tools.

Contribution

It introduces the concept of humans developing a theory of AI's mind (ToAIM) and empirically shows that brief training improves prediction of AI behavior in VQA tasks, challenging the effectiveness of existing interpretability methods.

Findings

Few-shot training (50 examples) improves human prediction of AI responses.

Existing interpretability tools like confidence scores and attention maps do not enhance prediction accuracy.

Humans can learn to anticipate AI failures with minimal training.

Abstract

Theory of Mind is the ability to attribute mental states (beliefs, intents, knowledge, perspectives, etc.) to others and recognize that these mental states may differ from one's own. Theory of Mind is critical to effective communication and to teams demonstrating higher collective performance. To effectively leverage the progress in Artificial Intelligence (AI) to make our lives more productive, it is important for humans and AI to work well together in a team. Traditionally, there has been much emphasis on research to make AI more accurate, and (to a lesser extent) on having it better understand human intentions, tendencies, beliefs, and contexts. The latter involves making AI more human-like and having it develop a theory of our minds. In this work, we argue that for human-AI teams to be effective, humans must also develop a theory of AI's mind (ToAIM) - get to know its strengths, weaknesses, beliefs, and quirks. We instantiate these ideas within the domain of Visual Question Answering (VQA). We find that using just a few examples (50), lay people can be trained to better predict responses and oncoming failures of a complex VQA model. We further evaluate the role existing explanation (or interpretability) modalities play in helping humans build ToAIM. Explainable AI has received considerable scientific and popular attention in recent times. Surprisingly, we find that having access to the model's internal states - its confidence in its top-k predictions, explicit or implicit attention maps which highlight regions in the image (and words in the question) the model is looking at (and listening to) while answering a question about an image - do not help people better predict its behavior.