Why your robot scares people — and what to do about it.

Your robot works. The engineering is solid. The specs are impressive. But when real people encounter it, they hesitate. They step back. They don't trust it. The problem isn't what your robot does. It's that your robot either makes weird sounds — or is silent.

The wrong fix

Most teams, when they notice the trust problem, reach for one of two solutions. Either they make the robot look friendlier — rounder shapes, softer colors, maybe cartoon eyes on a display. Or they add speech — a voice assistant that explains what it's doing.

Both miss the point. Friendly faces are cosmetic. They don't change how the robot feels when it moves toward you. And speech creates a conversation nobody asked for — a robot that narrates its every move is not reassuring, it's distracting. Meanwhile, replacing one generic beep with a slightly less annoying beep is not design either.

The actual problem is more fundamental: your robot has no sonic presence. Either it's silent — and therefore unreadable. Or it beeps — and therefore annoying. In clinical settings, this has a name: alarm fatigue. When every device sounds the same, people stop listening. Studies show that up to 95% of clinical alarms are false or clinically insignificant — and staff learn to ignore them all. In both cases, people can't intuitively understand what the product is doing, what it's about to do, or whether it knows they're there.

Per-sonare: why sound is how we read presence

The word "person" comes from the Latin per-sonare — to sound through. It originally described the voice carrying through the mask in Greek theatre. Before the audience could see the actor's face, they heard the character. Sound created presence before vision did.

This isn't just etymology. It's how humans are built. Hearing is the first sense to develop — a fetus recognizes its parents' voices months before it ever sees a face. And it remains the most immediate sense throughout life. You hear intent before you see it. You hear warmth, tension, attention, hesitation — all before you process a face or a gesture. When a colleague enters a room, you hear their footsteps, their breathing, the way they put down a bag. That's how you know whether to look up or stay focused.

A robot that beeps gives you noise. A silent robot gives you nothing. Neither moves with announcement. Neither greets. Neither confirms that it knows you're there. For the human brain, something that moves purposefully but doesn't communicate is a threat. And something that beeps incessantly is an irritant. Both fail at the same thing: making presence readable.

What non-speech communication actually means

The answer isn't adding beeps. Generic alert tones are the sound equivalent of cartoon eyes — superficial, and they can make things worse.

Non-speech communication means giving the robot a coherent sound language. A set of sonic signatures that communicate specific things:

The robot is about to move. The robot has seen you. The robot is waiting for you. The robot is working and shouldn't be interrupted. The robot is done and ready for the next step. Each of these states needs a distinct, recognizable sound — not a beep, but a designed auditory signal that humans can learn intuitively.

Think of it this way: a car communicates through sound constantly. The engine note tells you about speed and load. The turn signal tells you about intent. The parking sensor tells you about proximity. None of this is speech. All of it is communication. Your robot needs the same thing.

See also: Your AI has a voice but no character →

The operating room

We designed a non-verbal communication system for a surgical imaging robot. The original problem: the robot's audio cues were generic tones, indistinguishable from patient monitors in the operating room. Surgeons couldn't tell whether a sound came from the robot or from a heart rate monitor. The result was cognitive overload in a context where attention is life-critical.

The solution wasn't louder sounds or more sounds. It was distinct sound signatures — for movement types, radiation intensity levels, and system states. Each sound was designed to be immediately recognizable, even in a noisy OR. The surgeon could keep attention on the patient and still know exactly what the robot was doing.

That's what good product sound design does: it reduces cognitive load instead of adding to it.

Sound is a risk, not just an opportunity

Here's what most product teams don't realize: the wrong sound makes everything worse.

A generic beep from a sound library on a surgical robot sounds like a microwave timer. A cheerful jingle on an industrial robot arm feels condescending. An aggressive alert tone confirms exactly the fear you were trying to reduce. A constant hum that was meant to signal "I'm working" becomes the thing that drives people out of the room after twenty minutes.

Sound has the power to build trust instantly — but also to destroy it. This is why product sound is a design problem, not a technical one. The question isn't which frequency or which decibel level — it's what should the encounter with this product feel like?

Beyond feedback

Most teams, when they think about product sound at all, think about feedback: a tone that confirms an action, an alert that signals an error. This is the functional layer — necessary, but insufficient.

What builds trust is the layer above feedback: presence. The difference between a robot that suddenly moves and one that subtly signals "I'm about to move." The sound a system makes when it's idle — not silence, but a designed state of readiness. The acoustic equivalent of a person who looks at you and nods before speaking.

When a robot has both layers — functional feedback and designed presence — people stop backing away. Not because they understand the robot intellectually, but because they can read it. They know what it's doing, what it's about to do, and that it knows they're there.