WIRED's guide draws a line between finding a starting point and checking a fact. AI search may help you scan a broad topic. For an exact quote, a new update, or a high-stakes claim, open the original page. The guide offers advice, not a measured error rate for every search tool.
In a controlled preprint test, AI judges favored letters from some AI writing models. Some also favored letters made by their own model, even when candidate qualifications stayed the same. That is a warning for hiring-tool tests, not a measured rejection rate at real employers.
WIRED reports that researchers traced more than 400 linked apps and web domains in a network tied to abusive sexual deepfake tools. Apple said some approved apps later changed the websites they loaded. It removed 19 apps after WIRED contacted it. The report does not measure the whole app store.
In a randomized study, adults used an AI tool before medical visits. Clinicians judged more completed sessions to be well prepared with the structured tool than with a basic version of the same model. Patients also reported feeling more ready. The two measures used different groups of responses.
A research synthesis reports that people working with AI beat people working alone on average. Yet those teams did worse than the stronger of a person or AI working alone. Its latest stated run combines 1,955 measured results from 420 papers; the digest did not appraise each study.
A modeling preprint asks when AI systems that use tools should interrupt a person. The authors compare asking, finding a public answer, and continuing other work. Their tests include 864 synthetic cases, not measured time saved at a school or workplace.