Review AI outputs across everyday subjects — the entry route into Jobpeak at the Basic tier.
Tier: Basic (the entry route into the platform)
Generalist subcontractors review AI outputs across everyday subjects — current affairs, basic geography, common-sense reasoning, simple consumer math, and the kinds of plausible-but-false claims models often produce. The breadth is wide; the depth required at any one point is shallower than a specialist test.
Generalist is the entry route into Jobpeak. The work is also the work that exposes the most plausible-but-wrong claims, because the breadth is wider and the surface area for hallucination is larger. Calibration matters more here than in any other specialty — we'd rather a subcontractor say “I don't know, flagging for review” than guess confidently.
The Cognition Test for Generalist draws from a 150-question bank covering current affairs (date-checked), basic geography, household and consumer-science fundamentals, simple consumer math, and short-form reading comprehension.
10 easy / 20 medium / 10 hard in the domain section. Behavioural section unweighted by difficulty.
Date-checking on commonly-confused recent events; spotting model-generated answers that confidently cite non-existent sources; basic-geography sanity checks; consumer-math judgement (percentages, unit conversions, simple interest).
Calibration above all. We test whether you can distinguish what you actually know from what sounds familiar, and whether you flag the latter honestly.
Qualified Generalist subcontractors are matched to live AI review work within seven days. Rate ceiling is up to $30/hr at the Basic tier; the matching system surfaces work as it becomes available.
Many qualified Generalist subcontractors go on to take a STEM, Coding, or Professional-tier Cognition Test — there is no restriction on holding multiple qualifications. The Basic tier is a starting point, not a ceiling.
What you need to know about this specialty before you sit the test.