Mental health roles are now among the highest-advertised in AI training. On October 3, 2026 our feed held 37 live listings for psychologists, psychiatrists, counselors, social workers and related mental health professionals, from 7 platforms. 22 of them were posted in the five days before we looked, and they account for something striking: of the 76 listings in our entire feed that advertise a top rate of $200 an hour or more, 27 are psychology and psychiatry roles.
This guide covers what the work actually involves, who is hiring, what they ask for, and the caveats that matter before you apply. Every figure comes from the listings themselves. Where a platform's own listing text states a different rate from the one in its feed, we say so.
What AI companies want from mental health professionals
AI systems are increasingly used for conversations about mental health, and that is a domain where a fluent, warm and confident answer can be dangerously wrong. The listings describe three kinds of work, and they are quite different in size and commitment.
Judging simulated clinical conversations
The shortest tasks ask you to evaluate AI-generated conversations. Mercor's Clinical Mental Health Expert role asks you to review pairs of simulated clinical conversations and decide which one more closely reflects a realistic clinical interaction, applying your own clinical judgment. It is described as a short-term, fully remote task that takes up to 30 minutes. Terac's paid study for licensed psychologists, therapists and clinicians is larger: you review short AI conversations on mental health topics, write ideal responses and develop grading rubrics, over an 8 to 10 hour weekend block.
Writing and grading clinical scenarios
The most substantial work is authoring the tests that AI models are measured against.AfterQuery's Psychologist Expert role asks licensed psychologists to author realistic clinical vignettes, including risk assessment and ethical dilemmas, write reference responses showing the clinical reasoning behind each call, and grade model output for clinical accuracy, ethics and risk handling. One line of the listing captures the point of the whole field: it wants you to flag responses that are “warm, fluent, and wrong about risk,” and explain precisely why. Its social worker listing is similar, built around intake, safety planning, resource navigation and ethical dilemmas.
Challenging models and checking quality
Meridial's Psychology Specialist project asks you to challenge language models on topics like cognitive biases, therapeutic modalities such as CBT and DBT, psychological assessment tools and psychometric theory, documenting every failure mode. SME Careershas a quality assurance lead role for psychology projects, reviewing AI-generated content and other contributors' work for psychological accuracy, ethical awareness and diagnostic caution.
In the roles whose tasks we can read, the work is evaluating and shaping what a machine says, not treating patients or advising real people. It uses the judgment you already have.
Who is hiring, and what they pay
Here is every platform with a mental health listing, with the pay each one states. All rates are hourly unless noted.
| Platform | Listings | Stated hourly pay | The work |
|---|---|---|---|
| micro1 | 28 | $100–$200 (27 roles); $45–$95 (1 role) | Psychiatrist and psychologist roles by language, plus specialist roles. Open worldwide, no task description in our feed. |
| AfterQuery | 4 | Stated $60–$155 (psychologist $90–$155; psychology research $60–$120; social worker $70–$105; counselor $65–$100) | Author clinical scenarios and grade model output. Asks for licence plus two years' practice. |
| Mercor | 1 | $100 | Judge pairs of simulated clinical conversations. Up to 30 minutes. |
| Terac | 1 | $85 | Review AI mental health conversations, write responses and rubrics. 8–10 hour weekend block. Licensed clinicians. |
| SME Careers | 1 | Up to $80 | Quality assurance lead for psychology projects. United States. |
| Meridial | 1 | $8–$65 | Challenge models on psychology topics. United States. Rate set by experience and location. |
| Handshake AI | 1 | No rate listed | “Psychiatrists AI trainer”, via the Handshake AI Fellowship. Listed since May. |
A note on AfterQuery's figures. Its feed publishes a single midpoint for each role, so a role whose own text says “$90 to $155 an hour” appears elsewhere as $122.50. The ranges in the table are the ones stated in each listing's text, which is what you will be offered if you apply. The same caveat applies to our pay report, which counts what each platform publishes.
How does that compare with the rest of the market? Across all listings the median advertised top rate is $65/hr, medicine sits at $90/hr and science and research at $80/hr. Every platform in the table that states a rate gives a top rate at or above the market median, and micro1's $200 ceiling is one that only about 3% of priced listings in our feed reach. But the ceilings are not the offers: both micro1 and AfterQuery state ranges, and AfterQuery says plainly that your rate is set from your experience.
The micro1 language cluster
The biggest single group is on micro1: 22 listings, 11 for psychiatrists and 11 for psychologists, each tied to a language. They cover 11 languages: Cantonese, Georgian, Gujarati, Indonesian, Japanese, Lithuanian, Malay, Marathi, Russian, Tagalog and Vietnamese. All carry a stated range of $100 to $200 an hour, and they appeared in three batches on September 29, September 30 and October 3.
The pattern in the titles is consistent. Our classification marks the psychiatrist roles as requiring a licence, and the psychologist roles say PhD in the title. Every one is tied to a named language. micro1's listings carry no task description in our feed, so we cannot tell you exactly what the work involves. Read the listing before you apply, and treat the language as the defining requirement: these are roles for clinicians who work in a specific language.
Beyond the language batch, micro1 lists a Developmental Psychology (PhD) role, a Military/Veteran Mental Health expert role and a Child & Youth Abuse Trauma Specialist at the same $100 to $200 range, and a Substance Use, Adolescent Addiction Specialist role at $45 to $95.
What they ask for
- A licence, often. 15 of the 37 listings are marked as requiring a professional licence and 14 a doctorate. AfterQuery's Psychologist Expert listing asks for a PhD or PsyD with an active, unencumbered US licence and at least two years of clinical or research practice after licensure. Its social worker and counselor roles ask for the matching active licence and the same two years.
- Clinical depth, and clear writing. AfterQuery wants depth in areas such as assessment, psychotherapy or psychological testing, and the ability to write clinical reasoning clearly enough for another clinician to follow. The work is writing, not just judging.
- A relevant degree and clinical experience. Mercor welcomes backgrounds including clinical psychology, counselling psychology, mental health counselling and clinical social work.
- Language, for the micro1 batch. 22 of the 37 listings need a language other than English.
On location, 35 of the 37 listings state no country restriction in our feed. Read that carefully: a licence requirement can still be specific to a country. AfterQuery asks for an active US licence, for example, even though the listing has no country field. Two roles do name the United States explicitly: Meridial's Psychology Specialist project and SME Careers' quality assurance lead.
How the commitment differs
These roles range from a single 30 minute task to ongoing contracts, and that changes what they are worth to you. Mercor's evaluation task is a one-off. Terac's is a defined weekend block. AfterQuery's contract roles are described as remote and asynchronous with no fixed hours, a minimum of 10 hours a week and no maximum, paid weekly through Stripe, and the listing says there is no screen recording, keystroke monitoring or idle timers. If you want steady supplementary income rather than a one-off fee, the contract listings are the ones to look at.
Adjacent roles for allied health professionals
If you are a therapist but not a mental health clinician, AfterQuery also lists expert roles for occupational therapists ($70 to $105 an hour), physical therapists ($70 to $105 an hour) and respiratory therapists ($60 to $90 an hour), using the same build-and-grade-scenarios model. Our guide for healthcare workers covers the wider clinical field, and the doctors guide covers physicians.
Before you apply
- Check the confidentiality terms. You remain bound by your professional duties of confidentiality. Read each platform's terms on what you may use, and never bring real patient material into a task.
- Check your licence against the listing. Licence requirements are the most common reason an application goes nowhere, and they are often country-specific.
- Compare the stated range, not the ceiling. A range of $100 to $200 tells you what the platform will consider, not what you will be offered.
- Prepare for a screening step. Most platforms run an assessment or an AI-led interview first. Our interview guide covers what to expect.
- Apply only through the platform. No legitimate platform charges you to apply. Our scam guide lists the warning signs.
Common questions
Do I need to be licensed to do AI training work in mental health?
For most of the higher-paying roles, yes. 15 of the 37 listings require a licence and 14 a doctorate. Some roles, such as Meridial's Psychology Specialist project, are aimed at psychology specialists more broadly, so read the requirements on each.
How much can a psychologist earn doing AI training?
Stated top rates run from $65 to $200 an hour depending on the platform. micro1 states $100 to $200 for its psychologist and psychiatrist roles; AfterQuery states $90 to $155 for its Psychologist Expert role; Mercor and Terac state $100 and $85 for short evaluation work. These are advertised rates, not earnings, and your actual rate depends on your experience and on how many hours you are given.
Can I do it from outside the US?
Many listings state no country restriction, and micro1's language roles are built for clinicians in specific languages. But licence rules differ by country and some roles ask for a US licence, so check each listing. Our country pages show the roles that name your country, and our guide to working from outside the US covers the wider picture.
Is it legitimate?
The platforms in this guide are ones we track and review, but you should verify any platform yourself. Our legitimacy guide explains how, before you share personal or professional details.
See the live roles
Listings change daily, so treat this guide as a snapshot from October 3, 2026. For the current roles, see our medicine and healthcare and science and research pages, where these roles are filed, and the live pay report for how this field compares with the rest of the market.