Show HN: Interview Cheating Is on the Rise--Here’s How We’re Stopping It

[Vraj911]

Hi HN,

We’re a small team of engineers from Shivon AI building Lyra, an AI-powered hiring tool that helps companies identify truly qualified candidates while detecting interview cheating.

The hiring process today is frustrating—companies spend hours screening candidates, only to realize later that some weren’t the right fit. This problem has become worse with AI tools making it easier for candidates to cheat in interviews, leaving hiring teams unsure if they’re getting the right person.

We saw how much time companies waste on screening and how AI-assisted cheating is distorting the hiring landscape. We wanted to build something that ensures companies focus on the right candidates—without relying on outdated screening methods.

Lyra monitors candidates’ audio, video, and screen activities during the screening round to detect any suspicious behavior or AI-assisted cheating. It then flags inconsistencies in responses, giving recruiters valuable insights to make more informed decisions. Our goal is to bring transparency to the hiring process while saving companies time.

We’re still in beta and would love feedback from the community. If you’re interested, we’re offering limited beta seats for free—just let us know! You can check out more details <https://shivonai.com/> or review the demo <https://youtu.be/ZpvlJYDiRLk>.

Would love to hear your thoughts, suggestions, or any experiences you’ve had with this problem!

[erdaniels]

As a candidate, I would drop out of the interview process if I knew this was being used. This level of mistrust doesn't bode well for my future employer. As an employer, I don't care if someone cheats. In the context of programming and software engineering, I want to know that they explain the solution to the problem well and thoughtfully answer my questions. At the end of the day, we all cheat on the job!

[Vraj911]

I completely understand your perspective. The goal of Lyra isn’t to create an atmosphere of mistrust but to ensure a fair evaluation process. We’re not targeting code-related cheating, as our AI focuses only on verbal responses rather than programming tasks.

The challenge we’re addressing is the rise of AI-powered tools that feed candidates real-time answers during interviews, making it difficult to assess their actual understanding and communication skills. This goes beyond simple preparation—it allows candidates to present knowledge they don’t actually have, leading to mismatches between hiring expectations and on-the-job performance.

[threecheese]

What is your data collection policy with respect to a candidates voice, video likeness, and speech? Are candidates required to accept an EULA or similar acknowledgement wrt same?

[Shawnecy]

Why not just let candidates use AI like they would at work and design your interview test questions around that?

[laurent_du]

If you need AI to understand and answer interview questions, you are probably among the weaker candidates. Why would the company give you money? Why not hire someone who is smart and skilled enough to think by himself, and then if he uses AI to increase his productivity that's great! AI is not a replacement for your brain, otherwise why would we hire people at all?

[Vraj911]

Good point! AI is a great tool at work, but the issue is real-time AI assistance feeding answers during interviews. Lyra doesn’t aim to block AI entirely—just ensures candidates can explain concepts independently. Our AI asks verbal questions and diverse scenarios to estimate a candidate's potential.