ChatGPT resume builder: what it does well, what to switch off first, and what it can't check.
How to use ChatGPT, Claude or Gemini on a resume without handing over private data or inventing achievements: each assistant's own privacy and accuracy statements, a five-step workflow and prompts to copy.
ChatGPT can draft and tighten resume bullets, and OpenAI even runs a page called Resume Builder by ChatGPT, but it only knows what you tell it. Use it as an editor, not an author: remove your phone number, address and employer names you would not share, decide whether your chats may train the model, paste one role at a time, and check every line it returns against what you actually did. It cannot see how an employer's applicant tracking system is set up, so check the finished file separately.
ChatGPT for resumes, from OpenAI's own pages
Each row is OpenAI's own statement, read on the date in the sources.
| Topic | What OpenAI says |
|---|---|
| What OpenAI offers | OpenAI runs a page called "Resume Builder by ChatGPT" that says you can create a professional resume in minutes. |
| Who may see your chats | The same page notes that chats may be reviewed and used to improve OpenAI's AI models. |
| Training control | ChatGPT's data controls let you choose whether your conversations help train OpenAI models. |
| Default setting | On personal ChatGPT Free, Plus and Pro workspaces, data sharing is on by default; you can turn off "Improve the model for everyone" and keep your chat history. |
| Temporary chats | A ChatGPT temporary chat is not used to improve OpenAI models while it stays temporary, and OpenAI may keep a copy for up to 30 days for safety purposes. |
| Accuracy | OpenAI says ChatGPT can produce incorrect or misleading outputs and can sound confident even when it is wrong. |
Before you paste a resume: the setting in each assistant
A resume holds your name, contact details, employers and dates. Each cell is the company's own statement.
| Assistant | Model training, as stated | Temporary or incognito chats, as stated |
|---|---|---|
| ChatGPT | On personal ChatGPT Free, Plus and Pro workspaces, data sharing is on by default; you can turn off "Improve the model for everyone" and keep your chat history. | A ChatGPT temporary chat is not used to improve OpenAI models while it stays temporary, and OpenAI may keep a copy for up to 30 days for safety purposes. |
| Claude | For Claude Free, Pro and Max, Anthropic says it uses your chats to improve its models if you choose to allow it, or if a conversation is flagged for safety review. | Claude's incognito chats are not used to improve Claude, even if Model Improvement is turned on. |
| Gemini | Google asks Gemini users not to enter confidential information they wouldn't want a reviewer to see or Google to use to improve its services. | Turning off Gemini's Keep Activity setting stops future chats being reviewed to improve Google services, and temporary chats are not used to train Google's AI models. |
- ChatGPT's data controls let you choose whether your conversations help train OpenAI models.
- If you allow model training, Anthropic keeps that data for five years; if you don't, its 30-day retention period applies.
- Google asks Gemini users not to enter confidential information they wouldn't want a reviewer to see or Google to use to improve its services.
- Delete your street address, phone number and personal email from the copy you paste; add them back in your own file.
- Replace a confidential client or employer name with a description, such as a regional hospital network, if you would not post it publicly.
- Pick the training setting deliberately, or use a temporary or incognito chat for the resume session.
- Never paste references' contact details, salary history you have not chosen to share, or anything covered by a confidentiality agreement.
Why every line needs checking
- OpenAI says ChatGPT can produce incorrect or misleading outputs and can sound confident even when it is wrong.
- OpenAI lists fabricated quotes, studies and citations among the hallucinations ChatGPT can produce.
- OpenAI encourages users to approach ChatGPT critically and verify important information from reliable sources.
- Google says Gemini Apps may produce inaccurate or offensive information, including about people.
Advice: on a resume, the common failure is not a wrong date but a plausible achievement you never had, such as a percentage, a team size or a tool. Give the assistant your real numbers, tell it not to add any, and delete anything you could not explain in an interview. A resume match score from any tool reflects wording, not proof; see what a resume match score is.
A five-step resume session
- Prepare a clean copy. Your resume as plain text without contact details, plus one job description you are applying to.
- Ask for questions, not a rewrite. Have the assistant list what is missing or vague in each role compared with the job description, and answer those questions yourself.
- Rewrite one role at a time from your answers, keeping your verbs and numbers; reject any bullet that adds a fact you did not supply.
- Read it aloud. If a line does not sound like you, or you could not defend it in an interview, rewrite it by hand.
- Check the file, not the chat. Put the text back into your own document, save it as PDF or DOCX, and see what a parser reads.
A warehouse supervisor applying for operations coordinator roles
- Pastes his supervisor role (no contact details) and one coordinator posting, and asks which requirements his bullets do not show.
- The assistant asks about scheduling and inventory systems; he answers that he built the weekly rota for a team and ran cycle counts.
- He asks for two bullets using only those answers, deletes a suggested percentage he never measured, and keeps the rest.
- He saves the resume as a PDF in his own file and checks the parsed text before applying.
An invented person and situation, not a real customer or result.
The same discipline applies to tailoring in general: how to tailor a resume to a job description.
Prompts that keep the facts yours
Copy, then replace the brackets. They work the same way in ChatGPT, Claude or Gemini.
Find the gaps
Here is one role from my resume and a job description. List the requirements in the job description that my bullets do not show, as questions I can answer. Do not rewrite anything yet. [ROLE] [JOB DESCRIPTION]
Rewrite from my answers
Rewrite these bullets using only the facts in my answers below. Do not add numbers, tools, team sizes or results I did not give you. Mark any bullet where you had to guess. [BULLETS] [MY ANSWERS]
Cut to fit
Shorten this resume to fit one page for a [JOB TITLE] role. Remove the least relevant lines first and tell me what you removed and why. Change no facts. [RESUME TEXT]
More prompts for the rest of a search, from cover letters to negotiation: ChatGPT prompts for job search.
What a chat assistant can't do for a resume
- See the employer's system. No outside tool can see how a particular applicant tracking system is configured; check the file you send, then apply.
- Know what you did. It can only reorganise and phrase what you give it.
- Design and export reliably. Formatting from a chat window rarely survives a paste; use your own template or TryHired's free resume builder.
- Replace a second reader. A colleague in your field will spot claims that read as inflated; a resume review service is the paid version.
- Compare dedicated builders. If you want templates, saved versions and job tracking in one place, see AI resume builders compared.
Disclosure: TryHired makes HIRED, a paid job-search program, so we have a conflict of interest whenever we write about other products. Every product on this page is described from its own pages; HIRED is not a job board, a resume builder or a writing service.
ChatGPT resume questions
Can ChatGPT build a resume?
Is it safe to put my resume into ChatGPT?
Will recruiters know I used ChatGPT?
What is a good ChatGPT prompt for a resume?
Is Claude or Gemini different for resumes?
More on AI and resumes
How this page is built
Research-based, not hands-on: every product fact is taken from the company's own pages on the date shown and presented as that company's statement. We did not create accounts, buy plans or run tests, and we publish no ratings, scores, rankings or summaries of other people's reviews. Where a company's pages did not state something, this page says so instead of guessing. Sections marked Advice are a career coach's guidance.
- Whether an assistant's output is accurate for you: only you can check it against what you did.
- How a specific employer's applicant tracking system is configured: no outside tool can see it.
- Settings for work, team or enterprise accounts: this page covers personal plans only.
- Current prices, settings and terms: they change, so check the official page before you rely on them.
- OpenAI: chatgpt.com/writing/resume-building, fetched Nov 9, 2026.
- OpenAI: help.openai.com/en/articles/7730893-data-controls-in-chatgpt, read in Chrome (excerpt kept on file) Nov 9, 2026.
- OpenAI: help.openai.com/en/articles/8914046-temporary-chat-in-chatgpt, read in Chrome (excerpt kept on file) Nov 9, 2026.
- OpenAI: help.openai.com/en/articles/8313428-does-chatgpt-tell-the-truth, read in Chrome (excerpt kept on file) Nov 9, 2026.
- Anthropic: privacy.claude.com/en/articles/10023580-is-my-data-used-for-model-training, fetched Nov 9, 2026.
- Google: support.google.com/gemini/answer/13594961?hl=en, fetched Nov 9, 2026.
- Anthropic: www.anthropic.com/news/updates-to-our-consumer-terms, fetched Nov 9, 2026.
- TryHired: HIRED and free-tool facts from tryhired.co, read Nov 10, 2026.