In dystopian visions of the future, the humanoid takeover typically doesn’t look like this: a four-foot tall robot on wheels guiding you through the Baroque collection of an 18th century palazzo.
But for visitors to the Palazzo Madama, the House of Savoy’s old haunt in Turin, that’s precisely what they’ve been encountering in recent months. The robot in question is called R1, a pristine white thing with arms, an extendable torso, a bevy of cameras, and a battery life of two hours (at which point presumably museum fatigue sets in).
R1 has been programmed to tell visitors something of the story of how a noble family crossed the Alps in the 11th century, gradually established a powerful duchy, and would eventually offer up the first king of a unified Italy in the mid-19th century. It’s also primed to detail the paintings, tapestries, and furniture that decorate the former royal apartments. Visitors can address R1 directly and look into its beady LED eyes to ask it to go deeper or move along. This all takes place on Palazzo Madama’s first floor—R1 isn’t navigating stairs yet.
Dissonant as it might seem to be ushered around the prized collection of one Europe’s fallen royal families by an A.I.-powered robot, in some ways it’s no more than a physical embodiment of an experience already on offer at a range of cultural attractions, from Versailles to the Palais Populaire in Berlin. Nor is the Palazzo Madama the first to introduce a robot docent—one the institution will aid, rather than replace its human staff. Way back in 2018, the Smithsonian in Washington, D.C. deployed a four-foot-tall interactive robot named Pepper.
R1 inside the ceramic room of Palazzo Madama. Photo: Marco Bertorello/AFP.
The humanoid docent has been zooming around the Palazzo’s parquet floors (at a little more than one mile per hour) on-and-off for several years now as part of a project led by the Italian Institute of Technology (IIT) under Project Convince. Backed by €4 million ($4.7 million) in European Union funding, the goal has been to develop a robot that can autonomously adapt to its working environment and correct its mistakes. R1 has, in effect, been learning on the job.
For researchers keen to take R1 out of the laboratory and into the real world, the Palazzo Madama seemed an ideal testing ground. It’s a crowded, but calm environment with variable visitor flows. The space itself is large, weakly structured, and filled with valuable artworks, meaning R1 has to prioritize reliability and safety. An added bonus, the IIT noted, were the building’s “potential network connectivity issues” (i.e. the Wi-Fi is dodgy) leaving R1 to fend for itself. IIT reports that R1’s corrective software means it can now reliably roam the grand rooms all on its own — it completed 30 tours in December, 2025.
“The robot might get lost because it’s not able to localize itself in the same, but the system is able to automatically detects such an anomaly,” Lorenzo Natale, the coordinator of Project Convince said. “The robot will use its camera to look at the walls, relocalize itself, and then continue the tour successfully.”