Outlier vs DataAnnotation

Outlier and DataAnnotation both pay contractors for AI training work, and both keep their projects behind a login, so this site has no count of open roles for either. What you can read on their own pages differs. DataAnnotation publishes starting rates per hour by track and has one Starter Assessment. Outlier publishes one maximum, up to $150 an hour for ML/AI experts, and asks for an ID, a degree and a mobile number from the country you apply from. DataAnnotation suits someone who wants stated rates first and lives in a country its work is reported to cover. Outlier suits someone whose documents match its checks and who is fine seeing the tasking rate only after sign-up.

What is public and what is not

Neither platform has a public job board. Outlier shows projects inside your account after sign-up. DataAnnotation has role landing pages that describe kinds of work, and projects appear in your account after sign-up. That is why this page has no open-role counts or medians for them, unlike the pages for platforms that publish a feed. Everything below comes from their own FAQs and pages, from Trustpilot or from dated Reddit threads, and each source is named.

Listed pay on these pages is what the platform shows, not guaranteed pay and not an acceptance. A maximum such as "up to $150" describes the best case for one specialist role. A starting rate on a FAQ describes where a track begins, not what a typical worker earns.

Side by side

How the work and the pay work

Outlier is Scale AI's platform for contributors who write, rate and check AI training data. By its FAQ (read October 2026), rates vary with expertise, project complexity and location, and earnings are based on the hours you contribute each week. Payments are processed on Tuesdays for work done from the previous Tuesday to Monday midnight UTC. You work as an independent contractor, so taxes and insurance are yours. Available projects vary and may not match your field, and there is no promise of hours.

DataAnnotation pays contractors for writing, coding and fact checking tasks. Its FAQ (read October 2026) says there are no minimum hours and no daily login, that you choose your projects, and that work is not guaranteed. It also says low quality work can cost you access to projects. The FAQ does not name the company behind the platform, and press reports link it to Surge AI, which DataAnnotation has not confirmed on that page.

Getting in

For Outlier you create an account, verify your identity and pass assessments before you see projects. Its FAQ calls the process selective and says resumes are reviewed against the minimum requirements for specific domains, so signing up does not promise work. The ID, degree and mobile number must match the country you apply from.

For DataAnnotation you sign up, pick a track and take the Starter Assessment, then wait for approval by email, usually within a few days according to its FAQ. The FAQ names Persona as its identity verification partner. Our guide on how to pass AI training assessments covers what such tests check.

What to watch out for

• Expect gaps in work on both. Trustpilot reviewers and Reddit threads from August to October 2026 describe long stretches with no available DataAnnotation projects, and Outlier's own FAQ gives no promise of hours.
• Ask which hours are paid. Onboarding at Outlier and assessments at DataAnnotation may be unpaid.
• Accounts can be removed. The most common Outlier complaint on Trustpilot and Reddit is an account suspended or disabled without a clear reason, and some workers say earnings were held back. Single claims, which we cannot check.
• Lawsuits over contractor status were filed against Scale AI, which runs Outlier, in December 2024 and January 2025, and against Surge AI, reported as DataAnnotation's owner, in May 2025. They are allegations, not findings.
• Confirm rates, assessment format and accepted countries on the platform's own pages, since the DataAnnotation country list comes from third-party reviews.
• Never pay to work, and never buy, sell or hand over an account.

Verdict by type of person

• You want a published figure to compare: DataAnnotation shows starting rates by track, while Outlier shows a single maximum on its ML/AI Expert page.
• You have a Masters in ML or AI: Outlier's expert page lists up to $150/hr and a Masters requirement. That is a ceiling for one role, and the page gives no typical rate.
• You have no degree or no phone number from your country of residence: Outlier's FAQ asks for both, so check that before you start its sign-up checks.
• You live outside the US, Canada, the UK, Ireland, New Zealand or Australia: DataAnnotation publishes no country list, so check its pages before you spend an hour on the assessment.
• You want weekly payments on a fixed day: Outlier states Tuesdays. DataAnnotation states PayPal deposits within a few days of your request.

Platform pages: Outlier and DataAnnotation. Background on each: is Outlier legit, Outlier pay rates, Outlier AI on Reddit and is DataAnnotation legit. For rates across the platforms that publish roles, see the AI Training Pay Index.