Reporters & Correspondents

$40–75/hrrate as listed

micro1 recruits working reporters and correspondents to evaluate AI-generated news writing, source attribution, and factual claims against real reporting standards. Work is contract, remote, and paid hourly in an observed $40–75/hr band depending on beat, language, and assessed rubric accuracy.

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

What the work involves

You review model outputs that imitate journalism — wire-style briefs, explainers, interview summaries, headline sets, quote attributions — and judge them the way a desk editor would. Typical tasks include scoring two candidate drafts against a rubric and writing the rationale, flagging fabricated quotes or invented sources, checking whether a claim is actually supported by the linked material, and rewriting a passage to a gold-standard version that trainers can learn from. Some batches ask you to build prompts instead: construct realistic reporting scenarios (embargoes, anonymous sourcing, corrections, breaking-news uncertainty) that expose where models get things wrong.

The judgment being tested is editorial, not stylistic preference. Reviewers are expected to separate prose that reads well from prose that is verifiable, correctly hedged, and properly attributed — and to say plainly when a fluent draft is unpublishable. Written rationales matter as much as the scores; they are the actual training signal.

What the platform screens for

Verifiable bylines or staff/stringer history at a named outlet, wire service, or broadcaster
Depth under follow-up on a specific beat — politics, courts, business, health, science, sports, local government, foreign correspondence
Consistency against a rubric on calibration samples, including cases where the pleasant-sounding answer is the wrong one
Clear, fast written English; second-language reporting experience is a plus for localization batches

micro1's screening is AI-led: an asynchronous conversational interview probes your background and beat knowledge, followed by scored sample evaluations. Expect follow-ups that push past your first answer.

Logistics

Fully remote and asynchronous, with work drawn from queues rather than scheduled shifts. Most reviewers commit 10–20 hours a week; volume fluctuates with project cycles, so treat it as supplemental rather than steady. Contract engagement, hourly pay against approved time, with rates set per project and observed in the $40–75 range — higher end typically for niche beats, non-English work, or reviewers promoted to rubric-writing and quality audit.

What they look for

Document-based research
Source verification / fact-checking
Synthesizing lengthy documents
Narrative structure (lead
inverted pyramid)

Before you apply

Would you pass this screen?

Drawn from what this listing actually requires. Answered in your browser — nothing is sent to us, and we do not store your answers.

Have you had reporting or editing work published under your byline, or have you held a staff or stringer position at a named news outlet, wire service, or broadcaster?
Byline or staffing history · required
Can you identify at least one news beat you have reported or edited on for at least six months, where you developed specific knowledge of key sources, institutions, or patterns?
Depth on a specific beat · required
Are you able to write clear, grammatically correct English at the pace required to meet task deadlines?
Fluent written English · preferred

0 of 3 answered

Included with this listing

Interview prep — 11 questions

Matched to what this platform's screening actually measures. Answer each out loud in 60–90 seconds: claim → concrete example → implication.

What it tests: Whether you distinguish between a source's assertion and verified fact, and whether you'll push back on fluent prose that lacks grounding. This is the core editorial judgment the platform trains models to make.
A strong answer: A concrete example showing you checked a claim against source material (documents, interviews, records) and either included it with the right attribution or cut it despite good writing. The answer demonstrates you know the difference between 'the source said X' and 'X is true' and that you treat those differently.
Where people lose it: Describing a story where you trusted the reporting of someone else (a colleague, a wire service) without auditing the claim yourself, or conflating fact-checking with style revision. If you answer 'we rewrote it to sound clearer' instead of 'we couldn't verify it,' you've missed what the question is testing.