can an AI interview discriminate against you?
AI screening can be fairer than human screening, or it can scale the bias that was already there. The difference is what the platform was built to do. Here is the careful answer, with the questions that matter and what a serious platform does about them.
Tl;dr
Yes, badly built AI can discriminate. So can humans, often more. A serious AI platform is built to Remove the resume signals (Names, accents, gaps, formatting) that drive most screening bias, scores everyone with the Same rubric, And surfaces Evidence behind every score. The question is not AI or humans. The question is how the system was designed.
Where bias comes from in human screening
Before asking whether AI is biased, it helps to be honest about the baseline. Resume screening is one of the most studied parts of hiring, and the research is consistent. Resumes carry signals that have nothing to do with whether someone can do the job. Names, schools, address, employment gaps, formatting choices.
Recruiters skim each resume in seconds and pattern match. Different interviewers grade the same answer differently depending on their mood, the time of day, who they hired last week. These are not character flaws. They are the natural consequence of asking humans to make hundreds of fast judgments under pressure.
These are also the largest known sources of bias in hiring. Any conversation about AI fairness has to start there.
Where AI can help
A well designed AI interview removes several of the failure modes above. It is not a guarantee of fairness, but it changes the structure in important ways.
Structural advantages
Where AI can hurt
This is the part that deserves a real answer instead of a marketing one. AI can absolutely discriminate, and there are specific failure modes worth knowing about.
Real failure modes
These are real concerns and worth asking about. The difference between a fair platform and an unfair one is not whether AI is involved, it is whether the team building it took these failure modes seriously.
What a serious platform does
A serious platform answers each of those failure modes in design. An identical 15 minute structured behavioral interview for every candidate. Evaluation that does not see names, accents, gaps, or formatting. Scoring that traces back to specific moments in your responses, so a human can review the conversation record and check the work.
This is how aperture is built. It is also how any platform claiming to be fair should be built, regardless of brand.
How to tell if a platform is serious
You do not need to be a researcher to evaluate this. A few practical questions tell you most of what you need to know.
Signs of a serious platform
Your rights as a candidate
The rules vary by jurisdiction, but the questions you can ask do not. If a company is using AI to screen you, it is reasonable to ask how it works.
Questions you can ask
If you do not get a clear answer, that itself is information. Fairness is not a slogan, it is a set of practices. A company that cannot explain its practices in plain language is telling you something. One that can, even if the answer is imperfect, is showing you the work.
For the research behind this, including a real 2025 Stanford study on AI hiring discrimination, see Does AI hiring cause racial bias, and how does aperture avoid it.