LinkedIn AI interview screening just made automated candidate screening mainstream.
The feature is live for Hiring Pro subscribers: the AI conducts async audio or video screening calls with up to 40 candidates, generates a full transcript, and returns a 5-point rating based on how closely each candidate matches your pre-defined "ideal answers."
If you're on Hiring Pro, you may already have access. And at some point, a candidate you're evaluating will have gone through LinkedIn AI interview screening — whether you use it yourself or not.
Here's a clear-eyed look at what LinkedIn actually built — and where it falls short.
What LinkedIn's AI Interview Screening Actually Does
The workflow is straightforward. You set up an AI interview attached to a job posting. LinkedIn generates structured interview questions based on the role and recommends "ideal answers." You can edit both. Then you invite up to 40 applicants — or let LinkedIn select them — and each candidate completes an async audio or video screening call with the AI.
You receive back: a transcript, a recording, AI-generated summaries, and a 5-point rating.
That rating is generated by comparing candidate answers against the ideal answers you defined upfront.
It's a real feature, not a demo. It's available now. It's worth understanding before you turn it on.
What LinkedIn AI Interview Screening Gets Right
The underlying concept is sound, and a few elements deserve genuine credit.
Async video interviews over scheduled calls. Candidates complete screening on their schedule. Recruiters review results on their schedule. No calendar Tetris, no no-shows, no three rounds of rescheduling to get through a 20-minute first pass.
Structured questions tied to role qualifications. AI-generated questions tied to specific role requirements is a real improvement over improvised phone screens where first impressions depend entirely on which recruiter happened to pick up that day.
Richer output than call notes. Transcript plus summary plus rating is more comparable than a stack of handwritten post-call notes. You can actually evaluate candidates side by side using structured data rather than gut feel.
These are genuine improvements over the status quo for automated candidate screening.
Three Problems Worth Thinking Through Before You Deploy
1. The 5-Point Rating Measures Conformity, Not Quality
The rating is based on how closely a candidate's answers align with your pre-defined ideal answer. That sounds rigorous. It isn't — or at least, it's measuring something specific that you should understand.
What you're measuring is whether a candidate mirrors your expectations back to you. A candidate who gives a surprising but genuinely better answer will likely score lower than one who echoes your phrasing — since the AI is comparing answers against the "ideal answers" you defined upfront.
Good screening surfaces unexpected signal. This format rewards expected answers.
That's not a dealbreaker for every use case. But it is worth knowing exactly what you're evaluating before you let a rating drive your shortlist decisions.
2. The Legal Liability Is Yours — All of It
LinkedIn's documentation includes language stating that the hirer determines the lawful basis for processing for AI Interviews.
That's not boilerplate. That's LinkedIn explicitly transferring full legal responsibility to you for every data processing decision made during the AI interview.
LinkedIn also notes in its guidance that candidates may disclose health conditions or request accommodations during a screening call. If that happens, you are the data controller. You determine whether that processing is lawful. You own any regulatory exposure.
3. Candidates Notice When Their First Impression Is a Bot
This isn't a moral argument — it's a practical one about your employer brand.
The first interaction a candidate has with your company shapes how they feel about the process, the role, and the organization. A well-designed async assessment communicates that you've put thought into hiring. An audio call with an AI communicates something else — and what exactly depends on how it's framed, how the candidate is briefed, and whether they've encountered this format before.
AI screening is becoming more common. "Becoming normal" and "candidates feel good about it" are different claims. If you roll this out, watch your offer acceptance rate. That's your real signal.
What Purpose-Built AI Screening Looks Like Instead
LinkedIn AI interview screening is a first-generation integration of AI into a social network's hiring product. That context matters — it was designed to fit into an existing platform, not built from the ground up as a structured assessment system.
Screening designed specifically for this purpose starts from different assumptions.
It begins with the question: what signal does this role actually require? Communication clarity? Technical reasoning? Structured problem-solving? Each requires a different format. A text response reveals one thing. A video interview response reveals another. A coding challenge or practical exercise reveals something you simply cannot get from an audio call — regardless of how good the AI transcription is.
It treats candidate experience as an input to assessment design, not an afterthought. A structured 12-minute assessment tied clearly to real job requirements signals that a company respects a candidate's time. An open-ended audio call with an AI signals something different.
And it separates the AI's role from the hiring decision. In our view, AI's job is to surface signal from volume — to rank 50 submissions so a human can make a well-informed judgment call on the top 10. The AI doesn't make the hire. It makes the human's decision better.
When LinkedIn's Feature Is the Right Call
LinkedIn AI interview screening is a strong fit for a specific use case: lightweight, early-stage qualification for roles where you're sourcing exclusively on LinkedIn, candidate volume is under 40, and you need a simple pass/fail first filter.
If that describes your workflow, the feature is probably worth testing. The in-platform experience reduces friction for candidates who are already on LinkedIn, and it doesn't require adding a new tool to your stack.
For anything outside that use case — technical roles, volume hiring, roles where employer brand matters, or any situation where AI hiring compliance is a real concern — teams typically get better signal and more control from a purpose-built assessment platform. See our detailed comparison of LinkedIn vs. dedicated platforms →
Wonka combines text, video, and code assessments to surface genuine signal — not conformity. AI ranks candidates across communication clarity, technical depth, and problem-solving approach. No subscription, no seat fees — $1 per screened candidate. Run your first assessment free →