Software engineering jobs are under threat from AI. Some applicants are fighting back by using AI in the interview process, employing AI assistants that suggest responses on the fly during remote technical interviews.

Meanwhile, some employers are countering with—you guessed it—AI. They’re applying AI-powered tools to detect telltale signs of AI use during interviews.

This two-sided dynamic is turning hiring into an AI arms race with no clear winners. Yet as interviewers and interviewees navigate this daunting reality, experts believe the human aspect of the job search will prevail.

What’s driving the increase of AI in hiring?

AI hiring strategist Tatiana Teppoeva characterizes this phenomenon as playing cat and mouse in a climate of relentless AI-fueled tech layoffs and a job market filled with more applicants than open positions.

“What AI tools do well is identify if a person is performing according to some pattern or expected outcome,” Teppoeva says. When candidates experience constant rejection because they don’t fit the pattern, they might be forced to game the system using AI interview assistants, she adds.

Archie Payne, cofounder and president at technical recruiting firm CalTek Staffing, views it as a rational response to what he describes as a frustrating process from both sides. “Companies started to use AI resume screeners and similar tools to filter applications at scale. Candidates noticed this and started using AI in their interviews as a countermeasure to what they feel is a process that’s been automated against them,” he says.

This can lead to an AI-versus-AI loop, according to Ravi Kiran Pagidi, a senior AI data engineer at Navy Federal Credit Union who has been part of technical interview panels for software and data engineering positions. “The process may become less about actual capability and more about who can optimize better for the algorithm,” he says.

Tools of the trade

During technical interviews, software engineers might be tasked with outlining algorithms and answering questions related to system design and other software development fundamentals. Remote technical interviews usually turn into live programming sessions, with candidates writing code to solve a specific problem.

AI interview assistants such as Final Round AI, Interview Coder, and ParakeetAI can listen in, process the audio, and generate answers or code almost instantly. These tools can even be overlaid on the interview screen itself, claiming to appear invisible and undetectable.

“You’re able to read off an answer that’s coming to you in real time, so all you have to do is put on a little performance,” says Mudit Saraf, a software engineer at Meta.

Saraf and Shraddha Sunil, a software engineer at Microsoft, cofounded Ginger, an AI voice recruiter for first-round interviews. Ginger asks predefined questions and follow-up queries generated in real time, and it flags candidates who use AI during initial screening calls. The software tracks signals that include eye movement, a consistent delay in response times, tab switching, and speech patterns (phrases or sentence structures and flows) that “sound” like AI.

Sunil notes that Ginger has been tested mostly for entry-level roles for which applicants might be recent graduates or have only a few years of experience. “These candidates are more used to AI, and they use it a lot, so it’s nothing new to them,” she says.

Where AI hiring tools fall short

More employers are deploying AI-assisted interviewing platforms, Payne has noticed, with some seeing mixed results when it comes to AI detection. “The accuracy isn’t perfect yet in the platforms I’ve seen, and there have been a few times strong candidates were flagged as false positives,” he says. “That can be a serious problem when it can already be a challenge to find people qualified for the position without eliminating top performers for no reason.”

Teppoeva warns of other risks AI interviewing tools could pose, including privacy and security of applicant data, whether interview recordings will be used to train the models underpinning these tools, and bias and fairness.

A recent study from the Stanford Institute for Human-Centered AI, for instance, found that AI hiring tools can increase racial bias and give rise to systemic rejection. Following 3.4 million real job applicants, whose applications were all assessed by algorithms from a single vendor, the study found evidence of adverse impact for Asian and Black applicants.

These pitfalls highlight the need for human oversight. “I would definitely incorporate a human somewhere in the process and let humans have a say to make sure the results are fair,” says Teppoeva.

Audits, clear policies, and transparency are also a must for AI hiring tools, according to Pagidi. “Otherwise, qualified candidates may be filtered out unfairly, and companies may think they are improving efficiency while actually weakening the hiring signal,” he says.

Reasoning and authenticity go a long way

Instead of implementing AI detection tools, some tech companies including Meta are allowing AI use during technical interviews. AI-native software development platform Factory is treading the same path.

“We want our interview process to reflect how candidates actually do their jobs today using AI,” says Varin Nair, a software engineer who leads Factory’s technical hiring process. Applicants build a production-quality system or migrate a real codebase from one framework to another within an hour using AI coding agents. They’re then evaluated based on strategy rather than results.

“We explicitly do not grade on how many tests pass or whether they finished. We grade on planning, how they direct the AI, how they debug, and whether they can explain why their solution works,” Nair says.

He’s seen candidates surrender to an AI coding tool, accepting everything it returns. “AI is only as good as the judgement of the person using it,” says Nair. “Weak candidates lean on it to do their thinking and stall the moment it falls short, while strong candidates use it to move faster and free themselves to reason about architecture, trade-offs, and product.”

Such reasoning remains vital in software development. “Reasoning through edge cases and connecting the answer to production scenarios is where real engineering judgment shows up,” Pagidi says. “Developers will increasingly use AI tools, but they still need to own the final solution.”

CalTek’s Payne believes this approach of designing interviews to favor authenticity could benefit companies in the long run. “The best technical assessments I’ve seen lately are collaborative, involving codebase walk-throughs and architecture discussions in addition to coding,” he says. “It’s much harder to use AI to get through this kind of interview, so it’s a process that’s more likely to reveal how candidates really think.”

He also advises candidates to use AI to prepare but to keep answers their own during interviews. “Companies are getting better at detecting AI use, and getting caught can impact your long-term career prospects,” says Payne. “Technical communities are smaller than people think.” With each interview, applicants must weigh the risk and benefit of using these tools. Taking that risk, he says, rarely works in the candidate’s favor.