Why are interviews suddenly designed to catch AI?
Because employers now assume some candidates are outsourcing their answers, and a few are not even showing up as themselves. According to Semafor, reporting on 2026-08-07, AI avatars that mirror an applicant's voice and likeness are appearing in first-round job interviews; one Sydney recruiter, Lana Kersanava, said on record that she disqualified three such avatar candidates over roughly six months. That single account would matter little on its own, but it fits a broader pattern.
The Next Web, reporting in 2026-08 on Semafor and Greenhouse data, noted that nearly two-thirds of candidates now use AI in their applications, that 87% of US hiring managers use AI in hiring, and that AI-generated applications have surged by 239%. When screeners are automated on one side and answers are automated on the other, employers lose confidence that the person on the call is the person who will do the job. So they are changing the test.
The change is toward formats that are hard to fake in real time. The Next Web described a shift toward live technical walk-throughs, scenario challenges, and roleplay simulations. Alongside that, the Greenhouse 2025 AI in Hiring Report, which surveyed more than 4,100 respondents across the US, UK, Ireland, and Germany, found that 39% of US hiring managers are conducting more in-person interviews specifically to verify candidates, as independently reported by AOL and Business Insider in 2026-08. The friction is deliberate.
Does the soft job market make this worse?
Yes, because a slack market gives employers room to add scrutiny without losing applicants. When candidates are plentiful, hiring managers can afford extra verification steps that would feel excessive in a tight market. The US Bureau of Labor Statistics reported in its Employment Situation for July 2026, released 2026-08-07, that nonfarm payroll employment fell by 23,000 in July 2026 while the unemployment rate held at 4.1%. That backdrop means longer, more demanding interview processes are likely to stick around. Treat the extra hoops as the new normal rather than a temporary annoyance.
How do you prove you're the real candidate?
Make yourself easy to verify and hard to script. The goal is not to prove a negative about AI; it is to demonstrate a living, thinking professional whose knowledge is clearly load-bearing.
Start with presence. Turn your camera on and keep it on. Since some employers are wary of avatars, a steady video feed, natural eye contact, and small human moments such as reaching for a notebook or pausing to think all quietly establish that you are a person in a room. If your connection allows, avoid heavy virtual backgrounds and filters that can make a real face look synthetic.
Then show your reasoning, not just your conclusions. In a live technical walk-through, narrate what you are doing and why, including the dead ends. Say when you are unsure and how you would check. An AI-assisted answer tends to arrive polished and complete; a genuine problem-solver reveals the messy middle. Interviewers running scenario challenges are specifically looking for that middle, so give it to them.
Ground everything in specifics only you would know. When you describe past work, name the constraint you faced, the tradeoff you chose, the thing that broke, and what you learned. Generic best-practice language reads as retrievable and therefore suspect. A story about the migration that failed at 2 a.m. and how you rolled it back is not something a screener can generate for you.
Expect and welcome follow-up questions. The most reliable authenticity test is depth: a good interviewer will keep asking why until they hit either bedrock understanding or a bluff. Prepare by going one or two layers deeper on your own resume than the resume itself states. For every project you list, be ready to explain the decisions behind it, the alternatives you rejected, and what you would do differently now.
For roleplay simulations, practice thinking on your feet rather than memorizing answers. Rehearse the shape of common scenarios, an angry customer, a disagreement with a teammate, a missed deadline, then respond live to whatever variation the interviewer throws at you. Scripts crack under improvisation; frameworks bend and hold.
What should you avoid so you don't look like you're faking it?
Avoid anything that creates a lag between the question and a suspiciously perfect answer. Reading from a second screen, glancing repeatedly off-camera, or delivering fluent monologues with no hesitation can all read as coached even when you are simply nervous or well prepared. It is fine to pause, to ask for a moment, to request that a question be repeated. Those behaviors are distinctly human.
Do not over-polish. Candidates sometimes sand every rough edge off their delivery until it sounds machine-generated. A little spontaneity, an aside, a self-correction, a genuine laugh, works in your favor now.
Finally, be consistent across the whole process. If your written application is dramatically more articulate than your live conversation, that gap invites suspicion, especially given the 239% surge in AI-generated applications reported by The Next Web. Aim for a written self that a stranger would recognize when they meet you on camera. The safest position in this environment is to actually be the qualified person your materials describe, then let a live conversation confirm it.