A presenter said a team is building an AI "active priority engine" to forecast which experiments will have the biggest impact, so facilities spending large sums can focus on work most likely to advance scientific goals.