What it means
A prompt that gets a model to ignore its safety rules, for example through role-play, a hypothetical, or a harmful request split into harmless-looking parts. Red-teamers write jailbreaks so the model can be trained to resist them.
Used in: Red-teaming and safety roles.
In open roles right now
28 of the open roles we list mention “jailbreak” in the title or description, on 7 platforms (as of 23 October 2026). Of those that state an hourly rate (10 distinct postings), the median is $70/hr, and the middle half pay $50 to $90.
Most of them on: Alignerr (16), Handshake AI (5), Invisible (2), Weekday (2), Claru (1), SuperAnnotate (1), and 1 more.
AI Red Teamer (Remote)Handshake AI·United States·Full-time
$32–95/hr18 days ago
AI Safety and Policy EvaluatorTuring·United States·Flexible hours
Pay set per project29 days ago
AI Safety Experts - English & FinnishWeekday·United States
$48–62/hr1 month ago
AI QA Trainer - Freelance ProjectInvisible·Singapore·PhD
$6–65/hr6 months ago
Related terms
Adversarial promptA prompt written to make a model fail: to give a wrong, unsafe or rule-breaking answer.Red-teamingDeliberately trying to make a model produce harmful, unsafe or wrong output, so the developer can find and fix weaknesses before release. ← ID verificationPrevious term, A to Z. Labelling →Next term, A to Z. The whole glossaryEvery term on one page, A to Z.Open →