All insights

Free resource

AI Work Platform Research Checklist

A free, practical checklist for comparing AI training platforms, verifying opportunities, and recording eligibility, pay, privacy, and project terms.

4 min read

An attractive rate is not enough to evaluate an AI training platform. Before applying, you need to understand who operates it, whether your country and specialty are eligible, what work is actually available, how payment is calculated, and which personal information will be requested.

Use this free checklist for every platform or listing. Save one completed copy per opportunity so that you can compare facts instead of relying on memory or promotional claims.

1. Verify the organization and opening

  • I reached the opportunity through the organization's official domain.
  • The sender uses a verifiable company email address.
  • The role appears on an official careers, jobs, experts, or contributor page.
  • I checked the company name together with “scam,” “complaint,” and “review.”
  • I understand whether this is an active project, evergreen listing, talent pool, or job board listing.
  • No one asked me to pay for access, equipment, training, or release of earnings.

The Federal Trade Commission recommends verifying remote openings independently and warns that legitimate employers do not require upfront payment. Unexpected WhatsApp, Telegram, or text offers deserve particular caution.

2. Record eligibility before investing time

QuestionYour notes
Which countries are accepted?
Which languages are required?
Is professional experience required?
Are licenses, degrees, or identity documents required?
Is there an assessment?
Is this contractor, freelance, temporary, or employment work?
Does the platform currently support my payment country?

“Remote” does not always mean worldwide. A listing can be remote but limited by residence, tax status, language, client policy, or payment infrastructure.

3. Identify the actual work

Write down the task in plain language:

  • labeling or categorization;
  • response comparison or rating;
  • fact-checking or research;
  • writing prompts, reference answers, or rubrics;
  • language or localization review;
  • coding, mathematics, finance, medicine, law, or another specialty;
  • audio, image, video, data collection, or transcription.

Then ask what counts as accepted work. A high nominal rate matters less if instructions are unclear, reviews are unpredictable, or substantial time is unpaid.

4. Understand the payment claim

Record each figure exactly as the source presents it:

Payment fieldYour notes
Currency
Advertised rate or range
Hourly, per task, milestone, or salary
What time/output is billable
Payment schedule
Payment provider
Minimum payout
Platform or transfer fees
Tax responsibility
Dispute or rejected-work process

Do not convert an hourly project rate into an annual salary unless the opportunity guarantees those hours. Read AI Evaluator Pay in 2026 for the difference between published rates and effective income.

5. Review privacy and contract terms

  • I found the privacy notice and legal entity name.
  • I know why identity, tax, or biometric information is requested.
  • I know which provider processes payments and identity verification.
  • I reviewed confidentiality and intellectual-property clauses.
  • I understand whether my submissions may train models.
  • I know how to request account or data deletion.
  • I saved the terms that applied when I joined.

Only provide sensitive information inside a verified official process. A polished message or familiar logo is not proof of legitimacy.

6. Score the opportunity

Give each area 0, 1, or 2 points:

Area012
Identityunclearpartly verifiedofficial entity and domain verified
Role clarityvaguepartialtasks and acceptance rules explained
Eligibilityunknownsome conditionscountry, skills, and process clear
Paymentheadline onlypartial termscalculation, timing, fees, and disputes clear
Privacymissinggenericrelevant policies and processors identified
Riskasks for money/data earlyunresolved concernsno material red flags found

A score is a research aid, not a guarantee. New information can change the conclusion.

7. Keep an application log

Record the platform, official URL, role, country eligibility, application date, assessment stage, quoted rate, contact domain, and outcome. Never reuse confidential assessment content in your notes.

Before applying, learn how to become an AI evaluator and try an original task in the free Practice Lab. AI Work Prep is an independent educational resource and is not affiliated with the platforms you research.

AI Work Platform Research Checklist | AI Work Prep | Learn Practical Skills for AI Work