Zoom, Microsoft Teams, and Google Meet now auto-generate AI summaries and transcripts of job interviews by default on many accounts. Reporting published on September 18, 2026 warns that this creates a permanent, discoverable record of every interview β including any inappropriate or illegal question an interviewer asks, and any hiring decision an AI summary quietly biases. Over 90% of HR managers already use AI somewhere in recruiting, but most organizations still have no governance policy covering interview recordings. HR and recruiting teams should set consent, retention, and review rules before the next interview cycle, not after a complaint or lawsuit forces the issue.
What Is Changing About AI and Interview Recordings?
Video conferencing platforms used for interviews now default to recording and AI-summarizing conversations, turning what used to be an ephemeral conversation into a permanent, searchable transcript. That transcript captures both sides of the interview equally, including moments neither party expected to be preserved.
For years, an interview was a conversation both sides could reasonably expect would exist only as memory once the call ended. Auto-transcription and AI summarization remove that assumption entirely: every phrase, pause, and off-hand comment is now converted into searchable text and stored somewhere the organization may not have deliberately chosen, tied to a timestamp and a candidate’s name.
According to reporting from HR Dive published September 18, 2026, John Challenger, CEO of outplacement firm Challenger, Gray & Christmas, frames the risk in blunt terms: a candidate’s offhand mistake can now be recorded permanently, and an interviewer’s insensitive β or even illegal β question can be captured with equal permanence. Neither party controls what happens to that recording once it exists.
How Many Employers Are Already Using AI in Recruiting?
More than 90% of HR managers report actively using AI somewhere in their recruiting process, and nearly all of the remainder plan to adopt it within the next 18 months. Adoption has outpaced governance: four in five HR leaders say they are actively managing at least one unresolved problem caused by their AI tools.
That gap between adoption speed and policy maturity is the core issue. Recording and summarization features are frequently switched on by default at the platform level β inside Zoom, Microsoft Teams, and Google Meet β which means many hiring teams are generating this permanent record without having made a deliberate decision to do so.
Why Does an AI-Generated Summary Create Legal Risk?
An AI-generated interview summary becomes the de facto record a hiring manager reads instead of watching the full interview, and that summary can omit context, tone, or nuance that a human observer would have caught in real time.
Four Specific Risk Areas HR Teams Must Address
- Recording consent laws: Several U.S. states and international jurisdictions require one-party or two-party consent before a conversation can be legally recorded; interview platforms rarely confirm this automatically for every participant.
- Workplace monitoring and employment law compliance: A recorded interview can later be treated as evidence in a discrimination or wrongful-hiring claim, whether or not the organization intended to create that evidence.
- Bias amplification: AI summaries compress a 45-minute conversation into a few paragraphs, and that compression can strip out the context that would have explained an answer, increasing the risk that a biased or incomplete summary drives the hiring decision.
- Candor and interview validity: Knowing a conversation is recorded changes how both candidates and interviewers behave, which can distort the very signal the interview was designed to capture.
What Should HR Teams Do Before the Next Interview Cycle?
Given the compliance exposure, HR and recruiting leaders should treat interview-recording governance as an immediate policy gap rather than a future project.
1. Set an Explicit Consent Process
Confirm the consent requirements in every jurisdiction where interviews take place, and require an explicit, recorded confirmation from both the candidate and the interviewer before any recording or AI summarization begins.
2. Turn Off Default Recording Where It Is Not Needed
Audit the default settings on Zoom, Microsoft Teams, and Google Meet accounts used for hiring. Many organizations discover recording and transcription were switched on account-wide without a specific decision from HR or legal.
3. Require the Hiring Manager to Review the Original, Not Just the Summary
Make it a documented step in the hiring workflow that a decision-maker reviews the full recording or transcript for any candidate who advances past a first-round AI-summarized interview, rather than relying on the summary alone.
4. Set a Retention and Deletion Schedule
Define exactly how long interview recordings and AI summaries are retained, who can access them, and when they are deleted. An indefinite retention policy increases legal exposure without adding any hiring value after a decision is made.
5. Train Interviewers on What Not to Say on Record
Brief every interviewer that questions about age, family status, health, religion, or immigration status are now captured on a permanent, discoverable record β not just heard and forgotten β and reinforce which questions are legally off-limits before the interview happens.
How Should Consent Be Documented Across Different Jurisdictions?
Recording-consent law varies by state and country, and a single global template will not satisfy every jurisdiction an organization hires in. A defensible process treats consent as a per-interview checkpoint rather than a one-time clause buried in an offer letter or applicant portal.
| Consent Model | Where It Typically Applies | What HR Must Capture |
|---|---|---|
| One-party consent | Most U.S. states | Internal notice that recording will occur |
| Two-party / all-party consent | Select U.S. states and most EU countries | Explicit, affirmative consent from the candidate and every interviewer present |
| AI-specific disclosure | EU AI Act jurisdictions | Disclosure that an AI system is generating a summary or evaluation, separate from the recording consent itself |
Where hiring spans multiple regions, the safest default is to apply the strictest applicable standard β explicit two-party consent plus a separate AI-disclosure notice β to every interview rather than trying to vary the process by location, which introduces its own risk of inconsistent application.
What Happens if a Candidate Requests Their Interview Recording?
Once a recording or AI summary exists, it typically becomes a personal data record the candidate has a right to request under regulations such as GDPR or various U.S. state privacy laws, depending on jurisdiction. Organizations need a defined process for locating, reviewing, and producing that record within the applicable legal timeframe.
This is where an undefined retention policy becomes an operational problem rather than a theoretical one: if recordings are scattered across individual interviewers’ personal video-conferencing accounts instead of a centralized system, responding to a legitimate access request within the required window becomes difficult or impossible to do reliably.
Who Owns This Policy Inside the Organization?
Interview-recording governance sits at the intersection of HR, legal, and IT, and treating it as any single department’s responsibility usually leaves a gap. HR defines the hiring workflow, legal confirms jurisdiction-specific consent and retention rules, and IT controls the platform-level default settings that determine whether recording happens automatically in the first place.
Organizations already building AI governance frameworks under the EU AI Act should extend that same structure to cover recruiting tools specifically, since interview recording and candidate scoring both qualify as employment-related AI use cases under most current definitions.
Teams evaluating recruiting platforms with built-in AI features can review kurums.com’s comparison of AI recruiting software, and legal and compliance teams tracking the broader regulatory picture should reference the EU AI Act Article 50 transparency requirements that already apply to AI systems used in hiring decisions.
Frequently Asked Questions
Do video platforms record interviews by default?
Many Zoom, Microsoft Teams, and Google Meet accounts now enable AI transcription and summarization by default at the account or organization level, meaning recording can start without a specific decision made for that interview.
Can an AI interview summary be used as evidence in a hiring dispute?
Yes. Once a recording or AI-generated summary exists, it can be treated as documentary evidence in a discrimination, wrongful-hiring, or workplace-monitoring claim, regardless of whether the organization intended to create that record.
What is the single highest-priority fix for HR teams right now?
Confirming and documenting consent before any interview is recorded is the highest-priority fix, since missing consent creates legal exposure that no retention policy or review process can correct afterward.
Should hiring managers rely on the AI summary instead of watching the interview?
No. An AI summary should support, not replace, a hiring manager’s review of the full interview, because summaries can omit context and nuance that materially change how an answer should be interpreted.
Last Updated: September 19, 2026
Discover more from Kurums | Business Intelligence
Subscribe to get the latest posts sent to your email.