AI contractor qualification digest: separate test changes from rumor

Qualification talk travels fast in AI contracting. A changed screening question, a new test, or a résumé requirement can spread through a worker chat before anyone knows whether it applies to a particular role. For this edition, no platform-specific change to a qualification test, assessment rubric, résumé screen or onboarding standard is confirmed in the information available.

That is not evidence that requirements never change. It is a reason to separate a verifiable change from a repeated claim. For builders and contractors, the practical question is not whether an assessment feels different. It is what the platform has actually required for the specific role, and when.

Keep qualification claims in separate buckets

Three kinds of gate often get blended together in conversation. A platform may ask a candidate to qualify for access to a project; an application may screen for experience or availability; and an active project may apply ongoing quality checks. Those are different events. A report about one does not establish a new rule for the others.

The same distinction applies to prompts and tests. A task that asks for careful written judgment is not automatically a prompt-engineering assessment. A timed exercise is not automatically a platform-wide onboarding change. Unless the platform identifies the assessment and the role it covers, treat broad claims about a new standard as unverified.

Searches for phrases such as “DataAnnotation qualification test” or “Mercor assessment updates” can surface useful questions, but a search phrase is not proof of a recent change. Check the current role listing and any instructions delivered through the platform. Compare those with an earlier version only if you have one, and note the date and project. Do not assume a requirement on one project carries over to another.

Make your own application record

A small, factual log is more useful than trying to reconstruct an application from memory. Keep the platform and role name, the date you applied or tested, the requirements stated in the listing, and the result or status you received. If instructions change, record what changed in plain language. Do not save confidential prompts, answer keys or task content just to prove that a test happened.

For résumé screening, keep a copy of the listing and map each stated requirement to experience you can support. A past project may demonstrate relevant work, but it does not guarantee admission to a new one. Our earlier guide on matching résumé claims to verified project experience covers how to keep those claims specific without overstating them.

When evaluating advice from other contractors, ask for the scope: which platform, which role, when the person applied, and whether they saw a written requirement or inferred a change from the test itself. That does not make every report wrong. It helps distinguish an individual experience from a policy that applies broadly.

What a work record can and cannot show

Work history becomes relevant after qualification, when contractors need to describe their experience or reconcile what they worked. But a record of sessions, tasks and estimated earnings does not prove a test score, an application outcome or eligibility. Keep those records distinct.

RaterSidekick is a Chrome extension for tracking work sessions, tasks and estimated earnings across AI contractor platforms, including Outlier, Mercor, Handshake, DataAnnotation, micro1 and TELUS. It detects task completion with a user-selected keyboard shortcut and runs only on websites the user designates. It does not capture screenshots, keystrokes or task content. Users can also import history from spreadsheets, Clockify or platform exports, and its Pay Checker compares user records with platform reports. Those functions concern work records and pay comparisons, not qualification testing.

The useful move this month is modest: verify the requirement at the role level, preserve the date and wording of what you were told, and treat unconfirmed assessment chatter as a lead to check rather than a new rule. If a platform publishes a clear change, contractors can adjust to that. Until then, careful records beat guesses about what an unseen rubric might reward.