We found nine AI-related questions on seven employers’ application forms while capturing forms between September and October 2026, on Greenhouse, Ashby, Workable and other systems. The wordings below are quoted from those forms; employers are not named.
The four kinds of AI question on application forms.
| Kind | Wordings we captured | What a good answer gives |
|---|---|---|
| Read the AI policy | “AI Policy for Application” (a required yes or no) | Confirmation that you have read how the employer wants candidates to use AI. |
| Describe your use | “Describe how you use AI in your work” · “Which AI tools do you use in your workflow and for what?” · “Do you actively use AI tools in your work?” · “How do you get the most out of AI development tools?” | Concrete tasks, the tools you use for them, and how you check the results. |
| Show AI work | “Link any AI projects or open source contributions you’re proud of” · “Have you implemented AI workflows or AI-powered solutions in production or at customer sites?” | Evidence: a link, a shipped project, or a clear yes or no with one example. |
| Consent to AI notetakers | Interview recording and transcription by a named AI notetaking tool, with your approval | Your decision about being recorded in interviews. |
These sit next to the usual questions, often as required fields. The first and last are decisions; the middle two are short written answers that a hiring team reads alongside your resume.
Answering “How do you use AI in your work?”
Use four parts, one sentence each: the task, the tool, how you check the output, and what changed. The check is the part most answers leave out, and it is the part that shows judgment.
“I use an AI coding assistant to draft unit tests and refactor TypeScript, and an LLM API inside our document-processing service to extract fields from uploaded files. Every generated change goes through code review, and model output is validated against a schema before it reaches users. It cut the time I spend on boilerplate, which I now put into design reviews.”
“I use an AI assistant to summarise long reports and draft first versions of routine emails. I check every figure against the source and rewrite the tone before anything is sent. It gives me back about an hour on busy days for work that needs judgment.”
Replace the samples with your real tasks. If a time saving is a guess, say “about” or leave it out; an invented number is worse than none.
“Which AI tools…?” and yes or no versions.
- “Which AI tools do you use, and for what?” Pair each tool with a task. A short list of two or three is clearer than a long one.
- “Do you actively use AI tools in your work?” Answer truthfully. If the honest answer is no because of an employer policy, say so in any text field that follows.
- “Have you implemented AI workflows in production?” Count only work you shipped or ran for real users or customers. Prototypes are worth mentioning separately.
- “Link any AI projects you’re proud of.” Link the thing itself, such as a repository, pull request or write-up, and say what your part was.
AI policies you are asked to accept.
One AI company’s form includes a required “AI Policy for Application” question. It links to its guidelines for candidates and asks you to confirm you have read them by selecting Yes. The wording we captured invites candidates who work thoughtfully with its AI, while saying it wants to understand each candidate’s own skills and perspective through the hiring process.
Policies like this differ between employers and change over time. Read the linked text before you agree, and follow what it says about written answers, take-home tasks and interviews.
Consent to AI notetakers in interviews.
Two employers’ forms asked for permission to record and summarise interviews with a named AI notetaking tool, so interviewers can focus on the conversation instead of typing notes. Both framed it as an approval you give.
Decide before the interview, not during it. If you decline, it is reasonable to ask how notes will be taken. If you agree, speak as you would in any recorded call.
Should you use AI to write the answers?
Check the employer’s policy first; some ask for unassisted answers to particular questions. Where help is allowed, draft from your real experience and edit until every claim is true and specific to this role. Our guide on whether ATS can detect AI-written resumes explains what screening software does and does not check.
AI question answers people look for.
Name one or two real tasks, the kind of tool you use for each, how you check the output, and what changed as a result. Two to four sentences is enough. Avoid listing every tool you have tried.
Say so plainly and give the reason if there is one, such as a data policy at your employer. Then mention a specific way you would use it in this role, or how you have used it outside work, if that is true.
It asks you to confirm that you have read the employer’s rules or guidelines on using AI during the application. Read the linked policy before selecting Yes, because policies differ: some invite you to collaborate with AI while asking for your own perspective in written answers.
It is your choice. The forms we saw asked for approval and named the tool. If you prefer not to be recorded, decline and ask how notes will be taken; read the employer’s privacy notice if you want to know how recordings are stored.
Follow the employer’s stated policy first. If AI help is allowed, draft from your real experience, then edit until every claim is accurate and specific to this role. Generic answers are easy to spot whichever way they were written.