Human-AI collaboration is typically offered in one of two of user control levels: guidance, where the AI provides suggestions and the human makes the final decision, and delegation, where the AI acts autonomously within user-defined constraints. Systems that integrate both modes, common in robotic surgery or driving assistance, often overlook shifts in user preferences within a task in response to factors like evolving trust, decision complexity, and perceived control. In this work, we investigate how users dynamically switch between higher and lower levels of control during a sequential decision-making task. Using a hand-and-brain chess setup, participants either selected a piece and the AI decided how it moved (brain mode), or the AI selected a piece and the participant decided how it moved (hand mode). We collected over 400 mode-switching decisions from eight participants, along with gaze, emotional state, and subtask difficulty data. Statistical analysis revealed significant differences in gaze patterns and subtask complexity prior to a switch and in the quality of the subsequent move. Based on these results, we engineered behavioral and task-specific features to train a lightweight model that predicted control level switches ($F1 = 0.65$). The model performance suggests that real-time behavioral signals can serve as a complementary input alongside system-driven mode-switching mechanisms currently used. We complement our quantitative results with qualitative factors that influence switching including perceived AI ability, decision complexity, and level of control, identified from post-game interview analysis. The combined behavioral and modeling insights can help inform the design of shared autonomy systems that need dynamic, subtask-level control switches aligned with user intent and evolving task demands.

Recently, transformer-based language models have stretched notions of what is possible with the simple self-supervised objective of language modeling, becoming a fixture in state of the art language technologies [Vaswani et al., 2017, Devlin et al., 2019, Brown et al., 2020]. However, the black box nature of these models combined with the complexity of natural language makes it challenging to measure how accurately they represent the world state underlying the text. In order to better measure the extent to which these models can capture the world state underlying the symbolic data they consume, we propose training and studying transformer language models for the game of chess. Chess provides a simple, constrained, and deterministic domain where the exact world state is known. Chess games can also be transcribed exactly and unambiguously using chess notations (Section 2). Most importantly, the form of chess notations allows us to probe our language models for aspects of the board state using simple prompts (Section 3) and without changing the language modeling objective or introducing any new classifiers.

We conduct the first study of its kind to generate and evaluate vector representations for chess pieces. In particular, we uncover the latent structure of chess pieces and moves, as well as predict chess moves from chess positions. We share preliminary results which anticipate our ongoing work on a neural network architecture that learns these embeddings directly from supervised feedback. The fundamental challenge for machine learning based chess programs is to learn the mapping between chess positions and optimal moves [5, 3, 7]. A chess position is a description of where pieces are located on the chessboard. In learning, chess positions are typically represented as bitboard representations [1]. A bitboard is a 8 8 binary matrix, same dimensions as the chessboard, and each bitboard is associated with a particular piece type (e.g.