How to Use AI to Write Job Descriptions and Interview Questions Fairly
How to use AI to write job descriptions and interview questions fairly: prompts that start from the real work, how to spot biased or inflated requirements, building a structured interview with scoring guides, and the legal and privacy limits small employers should know.

Hiring is one of the most time-consuming jobs a small business owner or team lead does, and writing the job post and interview plan is often the part that gets rushed. AI chatbots can help a lot here. They can turn rough notes into a clear job description, suggest interview questions tied to the real work, and draft scoring guides so every candidate is judged the same way.
They can also copy the worst habits of job ads they learned from: inflated requirements, vague buzzwords and wording that quietly discourages good candidates. This guide shows how to use AI to write job descriptions and interview questions fairly, with prompts, a review checklist and the legal points a US employer should keep in mind.
Why fairness matters, beyond doing the right thing
A fair hiring process is also a better hiring process. Clear, honest job posts attract more qualified applicants and fewer mismatches. Structured interviews, where every candidate gets the same questions and is scored against the same criteria, make comparisons easier and decisions easier to explain.
There is also a legal side. Federal laws such as Title VII of the Civil Rights Act, the Americans with Disabilities Act (ADA) and the Age Discrimination in Employment Act (ADEA) apply to hiring no matter whether a person or a software tool wrote the materials. You are responsible for what you publish and ask.
Step 1: Start from the real work, not a template
The biggest mistake is asking an AI tool to "write a job description for a marketing coordinator." You will get a generic post full of requirements you do not actually need. Start with notes about the real job instead.
Before you open the chatbot, write down:
- The three to five outcomes this person must deliver in the first six months
- The tasks they will do every week, with rough time shares
- The tools and skills they truly need on day one versus what they can learn
- Pay range, schedule, location or remote rules, and benefits
- Who they report to and work with
Prompt example: job description from notes
You are helping a small business write a clear, inclusive job description. Use only the facts below. Do not add requirements I did not list. Keep it under 500 words, use plain language, and separate "Must have" from "Nice to have" (maximum 5 items each). Include the pay range and a short note inviting people to apply even if they do not meet every nice-to-have. Facts: [paste your notes].
The instruction "do not add requirements I did not list" matters. Without it, many chatbots add degrees, years of experience or skills that sound normal but are not needed for the job.
The AI prompt generator can help you structure prompts like this, and how to write better AI prompts explains why context and limits make such a difference.
Step 2: Review the draft for bias and inflation
Read every AI draft with a critical eye. Here is what to look for and how to fix it.
| Problem | Example | Fairer version |
|---|---|---|
| Inflated requirements | "Bachelor's degree and 5+ years required" for an entry role | List the skill instead: "Experience scheduling social posts for a business" |
| Unnecessary physical requirements | "Must be able to lift 50 lbs" for a desk job | Include only real duties, and note that reasonable accommodations are available |
| Age-coded language | "Digital native," "recent graduate," "young team" | "Comfortable with tools like Canva and Google Workspace" |
| Gender-coded or aggressive wording | "Rockstar," "ninja," "dominate" | Describe the actual work and results expected |
| Vague culture claims | "Work hard, play hard," "family" | Describe real schedules, flexibility and team habits |
| Too many must-haves | 12 requirements | 3 to 5 true must-haves; the rest as nice-to-haves or trainable |
Ask the AI to check its own draft
Review this job description for wording that could discourage qualified candidates, including age-coded, gender-coded or ability-related phrases, inflated requirements and jargon. List each issue, explain why, and suggest a replacement. Do not rewrite the whole post.
Treat the output as a second opinion, not a verdict. You still decide. AI tools can miss problems and flag things that are fine.
Check local requirements
Several US states and cities now require employers to include a pay range in job postings, and the details (company size, remote roles, what counts as a posting) differ. Check your state labor department's website before publishing. If you need to translate between hourly and yearly pay to write the range, the hourly to salary converter does the math.
Step 3: Build a structured interview with AI
Unstructured chats ("Tell me about yourself... so, what are your hobbies?") make it hard to compare people and easy to drift into topics you should not ask about. A structured interview fixes that.
Prompt example: interview questions tied to the job
Based on this job description, write 8 interview questions: 4 behavioral ("Tell me about a time...") and 4 practical or situational, each tied to one of the must-have skills. For each question, add what a strong, average and weak answer might include. Avoid questions about personal life, family, age, health, religion, national origin or other protected characteristics.
Turn answers into a scoring guide
Ask the AI to turn its notes into a simple 1-to-4 scoring rubric for each question. Then edit it with your own judgment. A small example:
| Question | 1 (weak) | 4 (strong) |
|---|---|---|
| "Tell me about a time you had to manage several deadlines at once." | General statement with no example | Specific example, how they prioritized, what happened and what they learned |
| "A customer emails angry that an order is late. What do you do first?" | Blames others or ignores the email | Responds quickly, checks the facts, offers options, follows up |
Use the same questions, in the same order, for every candidate. Take notes during the interview and score soon after, before comparing candidates. How to take better meeting notes has a format that works well here.
Questions to keep out of interviews
AI tools sometimes suggest "icebreakers" that drift into risky territory. Remove anything that asks about:
- Age, birth year or graduation year
- Marriage, children, pregnancy or childcare plans
- Health, disabilities or medications (you may ask whether the person can perform the essential job functions, with or without reasonable accommodation)
- Religion, national origin, citizenship beyond whether they are authorized to work in the US
- Arrest records or other topics restricted in your state or city
Step 4: Protect candidate privacy
Do not paste resumes, names, contact details or interview notes about real candidates into a consumer AI chatbot unless your tool's business terms and your own privacy policy allow it. Many free and personal plans may use conversations to improve their models unless you change a setting. Our guide on what not to share with AI chatbots covers this in more detail.
A safer pattern is to use AI for the materials (the job post, questions and rubric) and keep the decisions about real people in your own hands and systems.
Step 5: Be careful with AI that screens or ranks candidates
Writing a job post with AI is different from letting software screen, rank or reject applicants. Automated screening tools can repeat patterns from past hiring data, and some places now regulate them. For example, New York City requires bias audits and candidate notices for certain automated employment decision tools, and Illinois amended its Human Rights Act, effective January 1, 2026, to address discriminatory use of AI in employment decisions and require notice to employees. Other states are considering or phasing in similar rules.
If you use any tool that scores or filters candidates, ask the vendor how it was tested for bias, what notice you must give applicants, and how a human reviews its output. Before choosing one, use the checklist in how to evaluate an AI tool. This article is general education, not legal advice; an employment attorney can confirm what applies in your state.
A simple workflow you can reuse
- Write your own notes on outcomes, tasks, must-haves and pay.
- Ask AI for a draft job description using only those facts.
- Ask AI to review the draft for biased or inflated wording, then decide on edits yourself.
- Check pay transparency and other local posting rules.
- Ask AI for structured interview questions and a scoring guide; remove anything off-limits.
- Run every interview the same way and score before comparing.
- Keep candidate data out of tools that are not approved for it.
Frequently asked questions
Is it legal to use ChatGPT or another chatbot to write a job description?
Generally yes. Using AI to draft text is not prohibited, but you are responsible for what the final post says, including any discriminatory wording or required pay information in your state.
Can AI make my job description more inclusive?
It can help by flagging coded language, jargon and inflated requirements, but its suggestions are not always right. Review its notes yourself and keep the requirements tied to the real job.
Should I let AI pick which candidates to interview?
Be cautious. Automated screening tools can carry bias and are regulated in some places. If you use one, understand how it works, give any required notices and keep a human reviewing decisions.
How many interview questions should I prepare?
For a typical 45 to 60 minute interview, six to ten structured questions with time for follow-up and the candidate's own questions is a practical range.
Spotted an error? Tell us and we will review it under our editorial policy.



