Robots in the Mall: Public Attitudes Mapped

Robots invaded the CambridgeSide mall—and so did a real-world read on human trust.

In summer 2025, the RAI Institute staged a free, pop-up robot experience designed not just to show off hardware, but to understand how people actually feel about sharing spaces with machines. The setup split space into a museum-like lane where people could observe prototypes, and interactive corners where shoppers could try basic tasks with the robots—an approach that sounds romantic until you remember the hard math behind making a robot hold a door or avoid a startled passerby. The lineup prominently featured ANYmal, the quadruped from ANYrobotics, alongside a previous model of the RAI Institute’s UMV. The event was guided by Marc Raibert, the institute’s executive director and a familiar name in the robotics world, and it aimed to convert headlines about “AI sneaking into daily life” into grounded, experiential data about what works and what scares people in real spaces.

The intent was twofold: give the public a tangible encounter with machines that until now lived mostly in lab footage, and collect qualitative and observational data that could inform how such systems might be integrated into homes, offices, or factories without turning the public into a liability. The pop-up space was designed to surface the kind of friction that no white paper can predict—how a robot’s gait, speed, or blinking sensors are received when a person is standing in a mall corridor, not a sanitized test lab. Demonstration footage shows people leaning in, hesitating, or stepping closer with curiosity, as the robots perform guided tasks and respond to human cues.

From a technical standpoint, the event offers a window into a very real, very stubborn problem in humanoid robotics: public interaction at scale. The article notes the presence of two distinct platforms—ANYmal and a prior UMV model—highlighting a progression in physically embodied agents from a user-experience perspective. Yet the specifics that matter most to engineers—degrees of freedom, payload capacity, exact power and run time, and charging profiles—were not published for the pop-up. In other words, while attendees could judge how the machines looked and behaved in a public setting, the paper trail behind the hardware remains mostly hidden. For R&D teams, that silence is telling: it underscores a broader industry truth—the hard metrics that actually separate a demo from a deployable assistant are still tightly held, and often, not yet visible to the public.

What makes this event notable for engineers and investors is what it signals about readiness and perception. The TRL, or technology readiness level, of a mall demo is clearly not field-ready work; it’s a controlled, crowded-environment proof-of-concept. It’s the kind of exposure that can accelerate product-market fit by revealing what features people will tolerate and what they will not. It also surfaces practical constraints: in a shared space, a robot must be intentionally slow or predictable enough to avoid startling someone; it must gracefully handle the unpredictability of a busy corridor; it must communicate intent clearly to avert misunderstandings. These are the kinds of lessons that aren’t captured in a clean bench test but emerge when dozens of strangers interact with a device simultaneously.

Two practitioner insights emerge from this approach. First, public demos are powerful for shaping expectations—but they’re not a substitute for robust reliability engineering. The glamour of a smooth interaction in a mall can gloss over failure modes that only show up in real usage, like sensor occlusion in crowds or edge-case gait instability on uneven floors. Second, visibility matters as a lever for acceptance. If a broad swath of the public experiences a robot as approachable and helpful—without feeling surveilled or endangered—that sentiment can ripple into policy, funding, and ultimately adoption. But the opposite is equally true: a single tense encounter can fuel fear and resistance that outlasts months of clever demonstration.

As the industry buries its head in the next generation of demonstrations, the CambridgeSide popup serves as a reminder: the hardest gains come not from faster motors or fancier grippers, but from turning “demo-level” trust into long-term, everyday usability. The path from novelty to normalcy will hinge on transparent, shared progress—plus a willingness to publish the numbers that actually matter for real-world deployments.

Sources

Studying Human Attitudes Towards Robots Through Experience