Improved explanatory efficacy on human affect and workload through interactive process in artificial intelligence

Despite recent advances in the field of explainable artificial intelligence systems, a concrete quantitative measure for evaluating the usability of such systems is nonexistent. Ensuring the success of an explanatory interface in interacting with users requires a cyclic, symbiotic relationship between human and artificial intelligence. We, therefore, propose explanatory efficacy, a novel metric for evaluating the strength of the cyclic relationship the interface exhibits. Furthermore, in a user study, we evaluated the perceived affect and workload and recorded the EEG signals of our participants as they interacted with our custom-built, iterative explanatory interface to build personalized recommendation systems. We found that systems for perceptually driven iterative tasks with greater explanatory efficacy are characterized by statistically significant hemispheric differences in neural signals, indicating the feasibility of neural correlates as a measure of explanatory efficacy. These findings are beneficial for researchers who aim to study the circular ecosystem of the human-artificial intelligence partnership. Keywords: Affect; Brain Lateralization; EEG; Explanatory Efficacy; Human-centric Explainable Artificial Intelligence; Interactive Explanation; Workload 1. Introduction Recent advances in artificial intelligence (AI) and machine learning algorithms have resulted in models that not only achieve high predictive performance but also provide explanatory features to support their decisions, increasing model interpretability and transparency in real-world environments [1]. However, merely providing explanations is insufficient. Ultimately, AI should address the problems hindering human-agent interaction. Much of the current work for human-interpretable machine learning systems suffers from a lack of usability and efficacy [2]. Developing such a feedback-based interface for AI systems requires an evaluation on the strength of the cyclic relationship the interface exhibits, which we define as explanatory efficacy . Failing to integrate user knowledge with machine systems can decrease interaction quality to the point of causing interaction breakdowns. Consequently, the systems will lose their ability to justify their recommendations, decisions, or actions, resulting in a loss of trust from their users.