Emotion Detection in Installations: The Trend We Deliberately Don't Follow

Current trend reports on experience design keep circling back to one point: biometric installations that respond to heart rate or emotional state, plus emotion detection via cameras and facial analysis as the next stage for museums, trade shows, and brand spaces. The reasoning is understandable – the more precisely an installation knows how a visitor feels in the moment, the more precisely the experience can be tailored to it. From our own practice, though, I'm skeptical this is the right next step.

What PZ Nova does differently

Our interactive media facade PZ Nova in Pforzheim is controlled via depth sensors, not camera images. The difference is more than a technical detail: a depth sensor captures position and movement in space without ever storing, comparing, or interpreting a face anywhere. The facade knows that someone is standing there and how they're moving – it doesn't know who they are, and certainly not how that person feels right now. That decision was set from the start, not as a privacy measure bolted on afterward, but as part of the concept: an installation in public urban space, which thousands of people pass involuntarily every day, doesn't deserve technology that potentially knows more about them than they've agreed to reveal.

Why emotion detection promises more than it delivers

The actual catch with emotion detection isn't just privacy – it's the technology itself. Facial expression is an unreliable indicator of inner state – the same raised eyebrow can mean surprise, skepticism, or simply bright sunlight. Systems that try to derive reliable emotions from that are selling a precision that doesn't exist in that form. For an installation in public space, there's a further problem: whoever knows a camera is analyzing their face behaves differently than they otherwise would. The very unguarded curiosity a good installation is meant to spark gets dampened rather than heightened by the awareness of being watched.

The actual flaw in the trend's reasoning

The trend toward biometric and emotion-detecting installations implicitly assumes that more data about the visitor automatically leads to a better experience. That doesn't match what I've written elsewhere about experience design: the three principles that carry an installation in public space – instantly understandable, socially contagious, technically invisible – work precisely because they don't rely on knowing anything about the individual person. An installation that responds to position and movement can feel just as surprising and inviting as one that claims to read emotions – just without the risk that visitors feel watched instead of invited.

Where personalization still makes sense

That doesn't mean every form of personalization is wrong. In controlled environments with informed consent – a trade show booth where visitors deliberately register, an app that voluntarily shares data – personalized experiences make sense, because the person knows and decides what gets captured about them. Public urban space is exactly not that situation: nobody who happens to walk past a facade has agreed to being analyzed. That distinction – a controlled space with consent versus a public one without – should come before any decision about sensors, not the question of what's technically feasible.

Conclusion

Biometric sensors and emotion detection are currently being pitched as the logical next stage for experience design. From practice with PZ Nova, I see it differently: anonymous sensing, where it's sufficient, isn't a limitation – it's the technically and ethically more robust solution. The next step for good installations isn't knowing more about individual visitors. It's creating as much experience as possible with as little knowledge as necessary.