Review AI-generated privacy and compliance work products — DPIAs, records of processing, policy memos, training decks — and grade them against domain rubrics. Remote hourly work through Mercor, with observed rates of $80–120/hr depending on background and task type.

Summary, eligibility guidance and screening preparation are prepared by Sidequest from the platform's posting. Check the source listing before applying.

The work

You receive artifacts a model produced in response to a realistic privacy or regulatory compliance brief: a GDPR Article 30 record, a transfer impact assessment, a vendor DPA redline summary, a breach notification timeline, a slide deck training a sales team on CCPA opt-out mechanics. Your job is to decide whether a practitioner could actually use it. That means checking legal accuracy against the correct instrument and version, catching invented citations or misapplied thresholds, flagging reasoning that sounds fluent but collapses under scrutiny, and noting where formatting or structure undermines the document's purpose.

Feedback is written and structured. Rubrics typically separate factual correctness, analytical rigor, completeness against the prompt, and presentation quality — and reviewers are expected to score each dimension separately rather than issue a global impression. Comments should point to the specific span that is wrong and say what the correct treatment would be. Vague dissatisfaction is the most common reason evaluator work gets rejected.

What the screen looks for

Verifiable depth. Expect follow-ups on frameworks you claim: which articles, which state statutes, which sector regulator, what you actually did rather than what your team did.
Calibration. Can you distinguish a defensible-but-different approach from an error? Over-flagging stylistic preference as substantive failure is a disqualifier.
Deck and document fluency. Slides matter here more than in most legal evaluation work; you should be comfortable judging whether a PowerPoint or Google Slides deck is presentation-ready.
Writing. Your feedback is the deliverable, so the screen reads your prose closely.

Logistics

Fully remote and asynchronous. Work arrives in batches and volume fluctuates — some weeks offer 15–20 hours, others little. Most evaluators treat this as a supplement to practice or in-house work rather than a primary income source. Onboarding usually includes calibration tasks scored before you reach paid production volume.

Privacy / regulatory compliance Evaluator

The work

What the screen looks for

Logistics

Interview prep — 10 questions