Skip to main content
Open dataset · CC-BY-4.0 · Updated July 24, 2026

AI Interview Policy Tracker

Of the 20 companies tracked here, 10 ban it outright, 1 discourages it, 4 allow it, 4 require it, and 1 has no published rule.

The status split

Companies tracked
20

What counts as a policy here

Published company policies on candidates using AI during the hiring process (applications, assessments, live interviews). Candidate side only; employer-side AI screening is out of scope. Every row carries a verbatim quote, its source URL, the source date where the source is dated, and the date we last checked it. Companies without a verifiable published position are excluded rather than guessed.

status
The enforceable stance in the stage named by `stage`.
stage
The hiring stage the quoted rule applies to.
also
Nuance for other stages of the same company's process, when the source states it.
source_date
Publication date of the source. Null when the source page is undated; `last_checked` then carries the verification date.

For the argument behind the numbers, and what a candidate should actually do about them, read the full write-up.

Which companies ban AI in interviews?

10 of 20. A published rule prohibiting candidate AI use in the stage named. Most pair it with an explicit disqualification warning.

Anthropic

live interviews

This is all you–no AI assistance unless we indicate otherwise.

Also: Encourages AI when applying: create your first draft yourself, then use Claude to refine it. No AI in take-homes unless indicated.

Last checked

Goldman Sachs

interview process (all campus applicants)

As a reminder, Goldman Sachs prohibits the use of any external sources, including ChatGPT or Google search engine, during the interview process.

Also: Spokesperson: applies to all campus applicants across all positions. We want to hear from our applicants in their own voice.

Last checked

Accenture

assessments and interviews

Using AI to complete assessments (coding challenges, language tasks, etc.) unless explicitly stated otherwise

Also: Prohibited list also covers fabricated experience and voice cloning or deepfakes. Accenture states it uses proctoring tools and AI-detection methods.

Last checked

PwC

assessments (graduate applications)

While AI, including GenAI, can be useful in research, we tell candidates they should not use these tools during any assessment.

Also: PwC adds that activity undermining the integrity of recruitment will result in action.

Last checked

Deloitte

online assessments (UK early careers)

You'll also be asked to record a short integrity statement confirming your understanding and agreement of Deloitte's Integrity and AI use guidelines.

Also: Candidates record an integrity statement and must keep assessment content away from AI tools.

Last checked

KPMG

online assessments (UK graduate applications)

By completing our assessments, you are confirming that you are doing so independently and without the assistance of others, or any artificial intelligence or other resources, to provide responses that are not your own.

Also: Candidates found to have acted dishonestly, including using AI, are withdrawn and further action may be taken.

Last checked

Microsoft

assessments and interviews

During assessments and interviews, candidates should demonstrate their own skills without outside assistance unless explicitly permitted.

Also: The same page encourages responsible and ethical AI use during preparation, as long as it reflects true capabilities.

Last checked

Which companies discourage AI without banning it?

1 of 20. Written as an expectation rather than an enforceable prohibition, with no stated consequence.

Canonical

assessments and interviews

We ask that throughout your assessments and interviews you use your own words to describe your thoughts and ideas, without use of online aids, so that we can gather the best evidence we require to evaluate your suitability.

Also: Phrased as an expectation rather than a ban, and says online aids rather than naming AI.

Last checked

Which companies allow candidates to use AI?

4 of 20. Candidate AI use is permitted in the stage named, in some cases conditioned on disclosure.

Shopify

interviews, including coding interviews

You let them use whatever they want.

Also: Spoken description of practice by the VP of Engineering, not a formal written policy, made about director-level and above engineering hiring: candidates who do not use a copilot usually get creamed by someone who does.

Last checked

Meta

AI-enabled coding interviews (new interview type)

This is more representative of the developer environment that our future employees will work in.

Also: Reported from internal Meta communications seen by 404 Media describing a new AI-enabled coding interview type.

Last checked

Salesforce

interviews

The interview is your chance to show us how you think with AI, not just that you use it.

Also: Frames AI fluency as a signal evaluated across roles.

Last checked

GitLab

applications, take-homes and assessments, with disclosure

You may use AI to assist with your resume or cover letter, but the information must accurately represent your qualifications and skills.

Also: Live interviews: engage without reading from AI-produced answers. Take-home AI use must be cited and answered in your own words.

Last checked

Which companies require candidates to use AI?

4 of 20. The interview is built around AI use, so not using it is the deviation rather than the default.

Canva

technical interviews (Backend, Machine Learning and Frontend engineering)

we now expect Backend, Machine Learning and Frontend engineering candidates to use AI tools like Copilot, Cursor, and Claude during our technical interviews.

Also: An AI-Assisted Coding round replaced the computer-science fundamentals screen.

Last checked

Zapier

skills tests

In our revamped skills tests, we observe candidates working with AI in real time.

Also: Zapier states a bar that new hires must be AI-fluent.

Last checked

McKinsey

final-round case interviews (graduate hiring pilot)

In the pilot, some graduate candidates were asked to use McKinsey's AI tool, Lilli, during case interviews.

Also: Pilot-scoped, not all candidates: interviewers evaluated how applicants prompted the internal AI, assessed its outputs and applied conclusions to a client scenario.

Last checked

Coinbase

engineering interviews

come prepared to use the tools you'd use at work with us. We're not testing memorization or pattern recall; we're evaluating how you direct AI, how you critique its output, and how you exercise judgment to determine whether and when the model gets something wrong.

Also: Interviews rebuilt to evaluate how candidates direct AI, critique its output and catch its mistakes.

Last checked

Which companies have said something but published no rule?

1 of 20. On the record about candidate AI use without setting a rule either way. Included because the absence is itself the finding.

Google

interviews

We'll introduce at least one round of in-person interviews for people just to make sure the fundamentals are there.

Also: No explicit published ban. The stated response to AI-aided candidates is procedural: adding at least one in-person interview round.

Last checked

How to cite this dataset

Released under CC-BY-4.0. You may copy, republish and build on it, including commercially, as long as you credit the source. Use this line:

AI Interview Policy Tracker by UnchartedCareer, https://unchartedcareer.com/blog/companies-that-ban-ai-in-interviews, CC BY 4.0

Canonical page
https://unchartedcareer.com/research/ai-interview-policies
Version
schema v1.0, 20 rows, last updated July 24, 2026