The best AI detector for each job

One page per job. Each names a single pick, scored on the samples that job depends on, with false alarms on human writing counted and what the pick gets wrong said plainly.

Each pick comes from the same dated run as every other page on the site. A job page changes which samples count, never the weights: the pick for teachers is scored on student essays and the essays of writers working in a second language, the pick for free tools only on free plans, and the image pick on its own set of real photographs, real artwork and AI images.

Read the false alarm count before the score. A detector that catches every AI essay but flags a real student's work is the wrong tool for a classroom, and each pick page says plainly what its winner gets wrong. If your job is not listed, the overall ranking uses every sample, and the head-to-head comparisons set two tools you already know against each other.

Run RUN-2026-11-A started on 9 November 2026 and is in progress: GPTZero and Originality.ai are scored on the first 120 samples, and every other tool shows a test in progress tag until its readings are in.

How we testTest runs and raw files