Burger King is quietly turning its workers’ headsets into AI-powered hall monitors — and a lot of people think it feels less like a training tool and more like the start of a Thunderbolt of workplace surveillance.

The system has a friendly name: “Patty.” On paper, Patty is supposed to make life in a Burger King kitchen easier. It lives inside cloud-connected headsets, powered by an OpenAI model, and plugs into just about everything the restaurant does: drive-thru audio, kitchen equipment, inventory systems, and even customer feedback tools. Employees can ask Patty how many strips of bacon go on a Maple Bourbon BBQ Whopper, how to assemble a tricky limited-time burger, or how to clean the eternally cursed shake machine. It will also ping managers if the drink machine is running low on Diet Coke or if a QR-code complaint suggests the bathroom needs attention.

But that’s only half the story. The part that set social media on fire this week is Patty’s other job: listening in on employee conversations with customers and keeping score on how “friendly” they are. Burger King’s chief digital officer, Thibault Roux, says the company worked with franchise owners and customers to define measurable “friendliness.” The AI has been trained to recognize specific phrases like “welcome to Burger King,” “please,” and “thank you” in drive-thru interactions, then surface those patterns to managers as a kind of hospitality dashboard. In some pilots, Patty is also experimenting with judging tone — not just what workers say, but how they say it.

Inside Burger King, leadership insists this isn’t a Black Mirror episode. The company describes Patty as a coaching and operations tool, not a robotic boss scoring individual workers on politeness. A spokesperson has stressed that the system uses “aggregated keywords” to understand overall service patterns, and that it’s “not designed to track nor evaluate employees saying specific words or phrases” one by one. Managers, they say, can use Patty’s data to recognize teams delivering great hospitality, spot where service is slipping, and offer extra support or training where needed. The corporate line is clear: hospitality is “fundamentally human,” and the tech is just there to back up crews so they can stay present with guests.

Outside Burger King, the reception has been anything but polite. The Morning Brew post about Patty summed it up in one line: “Mind your manners. The King is listening,” and the comments immediately veered into “dystopian” territory. Some users joked about feeding Patty fake prompts; others flatly said they’d never eat there again. One person cut straight to the core critique: instead of wiring headsets with AI to count pleases and thank-yous, why not just pay workers more and create a better environment so they sound genuinely happier? That sentiment mirrored broader commentary from analysts and columnists who argue that when a low-wage job is already stressful, putting an always-on AI coach in your ear feels less like help and more like a threat.

Privacy advocates and labor experts are also raising the alarm. Patty is part of a growing trend where AI doesn’t just automate tasks, but evaluates human behavior in real time — from keystroke tracking in offices to cameras that assess warehouse workers’ efficiency. In Burger King’s case, Patty is listening to every drive-thru exchange, analyzing what’s said, and in some implementations, scoring interactions on friendliness metrics. Even if Burger King says the data is aggregated, critics note that it’s a short technical step from location-level insights to detailed individual profiles, especially when every worker is tied to a specific headset, shift pattern, and point-of-sale login. They argue that the line between “coaching” and “surveillance” is not defined by marketing copy, but by how the system is used over time — particularly if franchisees start tying Patty’s metrics to performance reviews, scheduling, or even firings.

There’s also a question of whether AI is even good at what Burger King wants it to do. Voice recognition systems still struggle with accents, background noise, and the controlled chaos of a fast-food drive-thru during a lunch rush. If Patty mishears a harried but polite worker as brusque, that could skew the data, especially in locations where staff are constantly multitasking between headsets, fryers, and in-person customers. And “friendliness” is notoriously hard to measure: a worker who skips the scripted greeting but nails accuracy, speed, and problem-solving might be more valuable to customers than one who parrots “you rule” with fake enthusiasm. Turning something as nuanced as hospitality into a handful of keywords and waveform patterns risks reducing customer service to a formula that doesn’t quite match what people actually care about.

For Burger King’s leadership, Patty is part of a bigger digital bet to fix a brand that’s had a rocky few years and lost ground to rivals. Alongside AI headsets, the company is tweaking its classic Whopper with a more premium bun and creamier mayonnaise, and pumping money into app upgrades and restaurant refreshes to win back customers. From their perspective, anything that can shave seconds off drive-thru times, keep machines running, and tighten up consistency across thousands of franchised locations is a competitive edge. The question is whether that edge is worth the cultural cost of being seen as the chain that literally wires its workers’ manners to an algorithm.

Zoom out, and Patty fits into a wider, uneasy moment for AI at work. Companies are racing to sprinkle AI into every corner of operations, from call centers to trucking fleets, often framing it as “assistive” tech while quietly pushing workers to adapt to new layers of monitoring. The debate around Burger King’s chatbot isn’t just about one fast-food chain, but about who gets to decide what “good behavior” looks like on the job — and who holds the data when that behavior is constantly measured. For now, Burger King says Patty’s rollout is a pilot, with hundreds of US restaurants testing the system and plans to expand access nationwide by the end of 2026. As those tests scale, the real story won’t just be how many times workers say “please” and “thank you,” but whether the people wearing the headsets feel like the AI in their ear is a helpful partner — or just one more boss they never got to vote for.