Topic
human-in-the-loop
Surprise-Guided MergeSort Reduces Human Ranking Costs by Using AI to Prioritise Comparisons
Researchers propose Surprise-Guided MergeSort (SGS), a framework that combines a Vision-Language Model with MergeSort to schedule pairwise comparisons for subjective ranking tasks. SGS routes only ambiguous comparisons to humans, achieving Kendall's τ×100 improvements of +6 to +12 over Active Elo under the same budget.
Technology Workers Spend Hours 'Botsitting' AI Each Week, Undermining Productivity Gains
New research from workplace AI firm Glean reveals that UK digital workers spend an average of 6.3 hours per week 'botsitting' – supervising AI outputs – negating half of the 12 hours they save through automation. While 78% of workers say AI makes them personally more productive, only 18% believe it significantly improves organisational performance. The report warns that traditional AI adoption metrics like seat count and prompt volume mask a growing quality-control problem, with over a third of AI sessions failing entirely.