arXiv preprintSeptember 15, 2026
Robin Welsch, Michelle Rausch, Daniela Fernandes, Pascal Knierim, et al.
Why it matters. In a 535-person study across 40 reasoning items, assisted participants beat unaided ones by roughly eight items, while the human and AI pairing did not clearly beat the assistant working alone. The behavior underneath is the finding: participants submitted the assistant's answer on 90.8% of trials when the advice was correct and on 74.3% when it was wrong. Almost all of the discrimination people think they are applying is not happening, and that 74.3% is exactly what an assessment should be built to catch.
MIT SloanSeptember 3, 2026
MIT Working Group on Generative AI and the Work of the Future
Why it matters. Interviews across more than 20 companies found AI improved jobs, rather than only speeding them up, where deployment choices were disciplined. The three failure modes give a test anyone can run on their own rollout: disuse, misuse, and overuse, held off by gathering evidence before scaling, teaching people when to trust, preserving teamwork, and keeping accountability on a named person. It reads as a checklist for the deployment and works just as well as a checklist for whoever you put in charge of it.
MIT SloanAugust 31, 2026
PNAS study by Sinan Aral and Harang Ju
Why it matters. A randomized experiment with more than 1,200 participants, about 1,100 ad raters, and roughly five million real-world impressions found the human and AI personality pairing changed both creative quality and market performance. Fit, task, interface, and interaction design decide whether the collaboration compounds, so "AI adoption" is too blunt a unit to measure anything by. Ask which person, on which task, with which setup, before crediting or blaming the model.
World Economic ForumSeptember 21, 2026
Shuvasish Sharma and the Global Future Council on Jobs and Frontier Technologies
Why it matters. WEF cites a 2025 study of 319 knowledge workers in which higher confidence in generative AI came with lower self-reported critical-thinking effort, while the thinking that remained shifted toward verification, integration, and stewardship. The role is changing shape rather than shrinking: context, expertise, feedback, decision rights, and capability-building stay with people. Those five are worth naming explicitly in the job description, because they are what the job now is.
