When AI Becomes the Workplace Lie Detector

Simon Carver

Hey everyone, welcome to the show! [warmly] I'm Simon Carver, and I'm joined today by Lachlan Reed and our brilliant guest, Jack Burns. Now, before we jump into today's deep dive, if you enjoy what we do here, please take a quick second to hit that like button, subscribe, and share the episode with someone who's navigating this wild, shifting world of work. Today, we're talking about something that sounds like it was ripped straight out of a Cold War thriller. For over a century, we've been obsessed with a single question: how do you know if someone is lying? It used to involve wires, rubber tubes, and a blood pressure cuff. But in 2026, the wires are completely gone. The interrogation room has dissolved. And the examiner... well, the examiner might just be a line of code. We are looking at the rise of the Algorithmic Inquisitor. Jack, you actually have a specific story that shows exactly how this is playing out right now, don't you?

Jack Burns

I do, Simon. [measured] Let's look at the case of Emily Chen. Emily wasn't a spy. She wasn't a criminal, nor was she trying to deceive anyone. She was a highly competent cybersecurity analyst interviewing for a remote role at a major technology company. The compensation was excellent, the hours flexible. A dream job. The interview process seemed standard. But right before the video call, she received a small, automated notice: "To ensure candidate authenticity, portions of this interview will be evaluated using advanced AI assessment technologies." Like most candidates, Emily clicked "accept" without thinking twice. We've been conditioned to accept automated resume scanners and video scoring. But what Emily didn't realize was that an AI model was measuring her. Thousands of data points per minute. Her blink rate, micro-expressions, voice cadence, pupil dilation, and facial blood flow.

Lachlan Reed

A blink rate? [scoffs] Blimey, Jack. So she's basically being targeted by a digital sniper who doesn't even need to pull a trigger to reject her. It's like trying to get a job while a speed camera clocks how many times your eyelids move.

Jack Burns

Precisely, Lachlan. [calm] The machine wasn't evaluating her cybersecurity skills. It was evaluating her baseline credibility. We have moved from a physical machine in a quiet room to an invisible, ubiquitous layer of software that operates through a standard webcam. The polygraph didn't die; it was digitized, scaled, and deployed into the cloud.

Simon Carver

And the wildest part is that when you're hooked up to an old-school polygraph, you *know* you're being tested. There's a giant rubber tube around your chest! But with this, it's just a tiny green light on your laptop bezel. It's completely invisible.

Simon Carver

To be fair to the companies building and buying these tools, there is a massive, genuine security headache driving this. We're in an era of absolute chaos when it comes to identity. Remote hiring has opened the door to incredible fraud. There are documented cases of synthetic identities, where someone uses a real-time deepfake over a webcam to pretend to be a credentialed software engineer, gets hired, and then downloads corporate intellectual property on day one. If you're a Chief Security Officer protecting billion-dollar source code, you are terrified. Humans are notoriously terrible at spotting these fakes. Studies show we perform barely better than a coin flip when trying to spot a liar. So, an AI that promises to scan biometrics and flag anomalies sounds like a godsend.

Jack Burns

The logic is sound on paper, Simon. [thoughtfully] But the systemic risk lies in where the boundary is drawn. Once these tools are introduced for hiring, the economic incentive is to expand them. We are seeing a quiet transition toward continuous behavioral monitoring inside the company. It's packaged as User and Entity Behavior Analytics, or UEBA. Software that runs in the background, analyzing changes in typing speed, communication style, file access patterns, and even sentiment in emails. The stated goal is cybersecurity and insider threat detection. But the psychological consequence is severe. It shifts the foundational workplace dynamic. It replaces the assumption of trust with a continuous audit of human intent.

Lachlan Reed

It's like fixing a classic motorcycle, right? [chuckles] If you're always checking the fuel line because you expect it to leak, you end up stripping the threads. You ruin the whole bike just by poking at it constantly. If you tell your team, "Hey, we trust you," but you've got a digital watchdog sniffing their digital exhaust every second of the day... guess what? They stop acting like trusted partners. They start acting like prisoners trying not to trip the alarm. You lose the very thing that makes a team great -- people taking risks and speaking their minds.

Jack Burns

Exactly. [matter-of-fact] When the environment becomes highly observed, behavior becomes entirely performative. People stop innovating because innovation requires deviation from the norm, and deviation is precisely what the algorithm flags as an anomaly. You optimize for compliance, but you destroy creative agency.

Lachlan Reed

And let's talk about the actual human cost here, because this is where my engine really starts to sputter. These algorithms are built on statistical averages. They assume a "normal" human response to stress. But what happens if you're neurodivergent? If you're on the autism spectrum, your eye contact patterns are completely different. If you have ADHD, your verbal cadence might jump around. Or what if you're a veteran with PTSD, or a survivor of workplace harassment who gets incredibly anxious just talking to a manager? The machine looks at a spike in facial blood flow or an irregular blink rate and flags it as "deceptive intent." It doesn't know you're just trying to process sensory overload. It has zero context.

Simon Carver

And then you get caught in this absolute nightmare of a feedback loop. Think about it: you sit down, you're told an AI is checking if you're a liar. Your heart rate immediately spikes. [nervous] "Oh god, what if it thinks I'm lying?" Your pupils dilate. Your breathing gets shallow. And those exact physiological reactions -- the ones caused by the *fear* of being falsely accused -- are the very signals the AI uses to clock you as deceptive!

Jack Burns

It is a classic self-reinforcing trap, Simon. [measured] The system creates the very stress it purports to detect, and then uses that stress to validate its own hypothesis. This is the exact reason why the United States Congress passed the Employee Polygraph Protection Act back in 1988. The law recognized that stress is not synonymous with guilt, and it strictly banned private employers from using lie detectors. Yet, tech vendors today bypass this by calling their tools "cognitive assessments" or "integrity analytics." We have effectively resurrected the 1988 polygraph, wrapped it in a modern software-as-a-service subscription, and branded it as a hiring optimizer.

Lachlan Reed

Spot on, Jack. [sighs] It's the old sheep in wolf's clothing... or wolf in sheep's clothing, rather. We're dodging the law just by changing the vocabulary.

Jack Burns

So where does this leave us? [thoughtfully] The solution is not to reject machine intelligence entirely. AI is highly effective at finding objective, technical anomalies -- like a login from an unexpected IP address, or a sudden, bulk download of proprietary data at three in the morning. Those are system-level anomalies. But we must establish a hard ethical boundary when it comes to translating physical, human biometrics into a verdict on human character or truthfulness. The machine should only ever be a diagnostic flag, never the final judge.

Lachlan Reed

Exactly! Let the machine point out the squeak in the engine, but let the human mechanic decide if the whole thing needs to be rebuilt. We can't let algorithms strip away our empathy and our common sense.

Simon Carver

Beautifully put, Lachlan. Ultimately, if we want to build workplaces that actually thrive, we have to protect the foundation of trust. Thank you so much, Jack, for joining us and bringing your deep, grounded perspective to this. This has been an incredibly eye-opening conversation. To everyone listening out there: keep your workforce human. Don't let the algorithms replace wisdom with mere data. If you loved this episode, please subscribe, leave us a review, and share it with a friend or colleague. Until next time, stay informed, stay human, and we'll see you on the next episode of The Human Workforce!