Home · Business · HR & People · Job ads & screening
As of 13 August 2026, AI can only partly match CVs to a job description.
This still needs a person who signs their name to it.
Can you do it?
5 minutesto a draft.
30 minutesto something you’d act on.
Cost, all in£0
Skill neededchat-fluent
Who has to check ita colleague
What the alternative costsThe alternative is manual screening by a hiring team; no price is provided here.
If this goes wrong: a suitable candidate is screened out or a protected characteristic is indirectly used, and your organisation carries the employment and data-handling consequences.
What to actually do
Hand it to a person
The route this page recommends
A person who owns the outcome does this end to end, worth it when the failure is dear.
Use a tool built for this
Second choiceDo it yourself
The distant thirdA chat interface, chat-fluent skill, and roughly 30 minutes until you can act on the result.
How to actually do it
- Open the current job description and copy it into a working document, marking each requirement as essential, desirable or unclear.
- Remove names, addresses, photographs, contact details and other unnecessary personal information from each CV, then assign each CV a neutral ID such as Candidate A.
- Check that you have a lawful reason and appropriate access to use the CVs for recruitment, and do not paste sensitive personal data that the comparison does not need.
- Paste the prompt followed by the job description and the anonymised CVs into ChatGPT, Claude or Gemini, then ask it to produce the evidence table and follow-up questions.
- Compare every assessment in the table with the exact CV wording and correct any claim that is unsupported, treating missing information as unknown rather than a rejection reason.
- Ask a colleague involved in the recruitment process to check the criteria, the shortlist and the reasons for excluding anyone before you contact candidates.
- Record the human decision and its job-related reasons, and use the output only as screening support rather than as an automatic hiring or rejection decision.
Prompt
Match these anonymised CVs to the job description below. Use only evidence written in each CV and do not infer age, sex, race, ethnicity, religion, disability, sexual orientation, pregnancy, marital status, nationality or any other protected or sensitive characteristic. Do not use names, addresses, photographs or gaps in employment as positive or negative evidence. First extract the essential and desirable criteria from the job description and flag any criterion that is vague, subjective or not job-relevant. Then create a table for each candidate with: candidate ID, evidence for each essential criterion, evidence for each desirable criterion, missing or unclear evidence, reasonable follow-up question, and an overall match assessment of strong, possible or weak. Do not reject anyone automatically and do not make a hiring decision. Treat missing evidence as unknown, not as proof that the candidate lacks the skill. Cite the CV section or wording supporting every assessment, state where the job description is ambiguous, and finish with a short list of checks I must complete before contacting candidates. This is not professional advice; for a serious employment-law or discrimination concern, consult an employment solicitor. Job description: [PASTE JOB DESCRIPTION] Candidate CVs: [PASTE ANONYMISED CVS WITH IDS]
Open it prefilled in ChatGPT or Claude, or copy it into Gemini, which takes no prefill link.
What it gets wrong
- AI cannot verify that a candidate's qualifications, employment history or achievements are genuine.
- AI cannot decide whether a vague requirement is genuinely necessary for the job or whether a different experience should count as equivalent.
- AI cannot reliably identify indirect discrimination created by the wording of the job description, the training examples or the shortlist.
- AI cannot take responsibility for rejecting a candidate or explain the decision as an accountable employer.
- Pasting full CVs into a general chatbot creates a personal-data handling decision that you still have to make and document.
What caps this at PARTLY: legal accountability, judgement under ambiguity and private data access.
How we scored this
Five axes, each scored nought to two by hand: ten means AI carries the task cleanly, and the thresholds that turn a total into YES, PARTLY or NO are published in the methodology. Each axis name links to its definition.
| Axis | Score (0–2) |
|---|---|
| Output | 2 |
| Inputs | 1 |
| Verification | 1 |
| Liability | 1 |
| Effort delta | 2 |
| Total | 7 / 10 |
The methodology and its thresholds are published in full.
FAQ
- Can AI screen CVs against a job description?
- Yes, AI can compare CV wording with stated criteria and produce a structured shortlist with supporting evidence. The answer is PARTLY because you must set lawful, job-relevant criteria, check the evidence and own the decision.
- Can AI rank candidates fairly?
- It can apply the same written criteria consistently, but consistent scoring is not proof of fairness. You must check for indirect discrimination, unsupported assumptions and criteria that are not genuinely needed for the role.
- Can I use ChatGPT to shortlist candidates?
- You can use it as a screening aid if your organisation permits it and you handle CV data appropriately. Do not let it make automatic rejections, infer protected characteristics or replace a documented human decision.
- Is AI CV screening legal in the UK?
- There is no simple yes or no: legality depends on how you use the system, the data involved, the recruitment criteria and the safeguards around decisions. This is not professional advice; consult an employment solicitor or data-protection professional for a serious case.
Nearby answers
Assessed by gpt-5.6-luna (gpt-5.6-luna) on 2026-08-13, second-checked by an independent model. Wrong somewhere? Email [email protected] and it gets re-checked.
The newsletter
AI news, new answers and product picks, straight to your inbox.