Human vigilance in the age of intelligent machines: Challenges and prospects

Ever-greater degrees of work are being undertaken by autonomous systems, and we are left to ask what roles humans will play in these coming, collaborative endeavours. Current consensus points to the increasing requirement for human supervision, largely in the form of sustained attention. The present work seeks to specify boundary conditions to such vigilance. It explores ways to assess the characterising decrement through evaluation of current measures, their shortfalls, and potential avenues for improvement. Decrement in human monitoring capacity is not an obligatory outcome in all hybrid work settings. The vigilance decrement is often avoidable through designs sensitive to the conditions of its occurrence. Where decrement remains, ameliorative recommendations are herewith proposed. The latter strategies are predicated upon improved measurement methods that feature the fuller integration of response accuracy and response latency. How these advances facilitate performance in human-AI teams forms the summary element of the present work.