Break ChatGPT, Earn $50 in Terac's AI Prompt Design Study
Terac is paying domain experts $50 to push ChatGPT until it fails on their own workflow. Here is what the "Domain Experts: Paid Assessment on AI Prompt Design" listing asks for and what the screening interview is like.
In August 2026 Terac ran a paid study called "Domain Experts: Paid Assessment on AI Prompt Design." It is no longer among the Terac studies we track as of October 1, 2026, but Terac reuses the same screening steps across studies, so this account still describes what applying looks like. The task: pick something from your own job, push ChatGPT until it fails to deliver a complete file, and get paid $50 for finding where it breaks. Getting considered starts with a screening interview that takes about three minutes.
Here is what the listing asks for, and what the screening interview is like.
What the listing says
- $50 one-time, remote, about an hour of work. You build a demanding prompt from a real workflow you know deeply, test it in ChatGPT, and refine it until the model fails to deliver a complete file.
- The whole task is recorded: screen, camera, and microphone, while you think out loud. You hand back your prompt, your failure notes, a grading rubric, and the file ChatGPT generated.
- The screening interview itself took three minutes and also asked for camera and screen share, before the paid task ever starts.
- Not everyone qualifies. The listing rules out H1-B and STEM OPT candidates, and pays only once your work is approved.
- High performers get considered for ongoing paid research projects, not just this one session.
Latest Terac Opportunities
Showing the 6 latest open roles
Software Engineers: Paid Code Review for AI Agent Evaluation
$60-80
/hr
Photography Evaluators: Paid Task on Image Quality and Brand Guidelines
$50-70
/hr
South Asian Software Engineers: Coding Tasks for AI Evaluation
$30-40
/hr
Licensed Mental Health Clinicians: AI Conversation Review
$80-100
/hr
What the listing asks domain experts to do
Terac's own framing, from the listing: they are running a paid study on advanced AI capabilities to find people who excel at designing tasks that expose where models fall short. It feeds into a larger effort to evaluate and improve how AI handles complex, multi-step requests, and high performers in this trial get considered for ongoing paid research projects, not a one-off payout.
The listing is open to professionals, domain experts, and knowledge workers from any field who deeply understand a complex, specialized workflow, the kind of person who can recognize a high-quality output in their own area within seconds. The real filter is whether you can spot a wrong answer in your field faster than a generalist reviewer could.
What the paid hour involves
Straight from the listing, this is the sequence:
- Translate a familiar, complex workflow from your job or personal life into a demanding prompt for ChatGPT.
- Test that prompt in ChatGPT and identify where it fails to produce a complete file.
- Record your screen, camera, and microphone while you think through the task out loud.
- Refine the prompt to make it harder, then write a clear, structured rubric a stranger could use to grade any AI's attempt at the task.
- Hand back your prompt, your failure notes, the rubric, and the file ChatGPT generated.
The "complete file" detail matters. The task is built around ChatGPT producing a deliverable, a spreadsheet, a document, a piece of code, and finding where that output breaks down. Whatever workflow you pick should end in a real file rather than a paragraph of text.
What the screening interview asked
The AI interviewer opens by asking about your background and how you use AI tools day to day. It wants specific examples, not a summary of your job title. If you already use AI tools regularly for real work, this part is easy, you are just describing what you already do.
It moves quickly and stays adaptive, following up on what you just said rather than reading down a fixed script. The whole thing runs about three minutes, closer to a quick screen than a real conversation, but a smooth one. It also asks for camera and screen share during the interview itself, not only for the paid task later, which fits: if the eventual task involves recording your screen, camera, and microphone, checking you can and will do that at the interview stage works as a filter.
The contract terms worth reading before you apply
The listing spells these out plainly, and they are worth knowing before you spend the hour:
- You are engaged as an independent contractor, fully remote, on your own schedule.
- The opportunity can be extended, shortened, or concluded early depending on needs and performance.
- Your work will not involve access to confidential or proprietary information from any employer, client, or institution, you are drawing on your own knowledge, not leaking anyone else's data.
- Payment is issued once your work is approved.
- Terac cannot support H1-B or STEM OPT candidates for this opportunity.
What happens after you finish the interview
Finishing the interview does not hand you the paid task. Applying runs through three steps: connecting your profiles, ID verification, and the screening interview itself, and Terac's own note is that these steps are reused across roles, so you should not have to repeat the same interview twice for a different study.
A team reviews applicants after that and emails the people it selects, usually within a day or two. Finishing the interview only means being considered, not being invited to the paid task itself.
Know someone who fits this better than you do?
Terac runs its own referral program, with the terms shown in your Terac dashboard.
Terac's own bar for a good referral: someone who uses AI tools for work, has deep knowledge of one specific workflow from their job or personal life, has access to a laptop or desktop, and is comfortable sharing a screen recording with audio and video during the task. If you know someone like that, an accountant who lives in spreadsheets, an analyst who writes reports, a developer, this is a better fit for them than for a generalist.
Worth the three minutes?
Yes, if you already use AI tools for real work, have a workflow that ends in a real file, and do not mind being on camera and mic for an hour. The screening interview costs almost nothing, three minutes and a webcam, and the listing is upfront about pay, terms, and who cannot apply.
This covers the listing and the screening interview. The paid task itself starts after acceptance.
Think you or someone you know fits this study?
Deep knowledge of one workflow, a laptop, an active ChatGPT account, and comfort being recorded on screen, camera, and mic for an hour.
Take the Terac Screening Interview →Affiliate link: we may earn a referral fee at no cost to you. See our disclosure.
FAQ
What is the real title of this Terac study?
"Domain Experts: Paid Assessment on AI Prompt Design." It pays $50 one-time and is listed as a remote, one-time opportunity. "Break ChatGPT on your own workflow" is how the brief describes the task itself, not the listing title.
How long does the paid task take?
About an hour, according to the listing. You build a demanding prompt from a real workflow, test it in ChatGPT, refine it until the model fails to deliver a complete file, then write a grading rubric, all in one screen, camera, and microphone recorded session.
How long does the Terac screening interview take?
About three minutes. It is a short, adaptive conversation, not a form. It asks about your work and how you use AI in it, and moves fast if you already use AI tools regularly.
Does Terac record your camera and microphone, not just your screen?
Yes, for the paid task itself, all three: screen, camera, and microphone, while you think through the task out loud. The screening interview also asks for camera and screen share.
Can H1-B or STEM OPT candidates apply?
No. The listing states Terac cannot support H1-B or STEM OPT candidates for this opportunity.
Do you get the paid study right after the interview?
No. The interview only gets you considered. Terac reviews applicants and emails the people it selects, usually within a day or two. Finishing the interview is not the same as being invited to the paid task.
Related guides
Terac review: full platform breakdown covering pay rates, how matching works, and who qualifies.
Browse live Terac job listings: current studies updated regularly.
Terac platform intel: active projects, pay data, and expert network reports.
Cite this page
"Break ChatGPT, Earn $50 in Terac's AI Prompt Design Study", aitrainer.work, aitrainer.work/guides/terac-ai-prompt-design-interview

MSc Human-Computer Interaction | Founder
Pietro is the founder and technical lead of aitrainer.work. He builds and maintains the platform's data pipeline, certification infrastructure, and editorial standards.