At Milan’s Pinacoteca di Brera, artificial intelligence has been given a focused museum task: recognise one of eight paintings and deliver accessibility content created by people. The new Sensing Art—Masterpieces Beyond Sight pilot uses AI glasses to help blind and low-vision visitors navigate the collection with greater independence.
The project was developed by Meta, the Pinacoteca di Brera and EssilorLuxottica with the National Association for the Visually Impaired, the Italian Union of Blind and Partially Sighted People and Milan’s Institute for the Blind Foundation. It extends DescriVedendo, an accessibility programme that has operated at the museum for almost a decade.
The model recognises; the museum interprets
When a visitor faces one of the eight works, the glasses identify it and play the corresponding audio through open-ear speakers. Each work has three layers of material: a morphological description, historical and artistic context, and an artist biography. Short orientation tracks support movement between works.
The crucial design choice is that the interpretive language is not generated spontaneously by the model. It comes from the museum’s existing programme, created for blind and low-vision audiences using the DescriVedendo method. AI performs recognition and delivery; human specialists remain responsible for the account of the artwork.
That boundary makes the pilot more consequential than a generic product demonstration. Museum interpretation involves judgment, language, scholarship and accountability. By separating machine perception from authored content, Brera creates a clearer chain of responsibility when a description is incomplete, inaccurate or insensitive.
Access as infrastructure
The glasses are available free of charge by reservation from Tuesday to Friday, 10:00 a.m. to 4:30 p.m. The programme also integrates orientation support, including access to Be My Eyes, rather than assuming artwork recognition alone creates an accessible visit.
There are limits. The pilot covers only eight works, the hardware model has not been publicly specified, and no end date or visitor data has been announced. These gaps matter because accessibility should be evaluated by reliability, availability and user experience—not by novelty.
What this means for artists and museums
For institutions, Brera offers a practical principle: use AI to remove a specific barrier, then surround it with human-authored content, user testing and non-AI support. The technology is most credible here when it is one layer in a mature accessibility practice.
For artists, the project expands the meaning of an exhibition-ready work. Descriptive language, orientation cues and accessible interpretation can be considered during production rather than added after installation. Artists can document materials, spatial relationships and intended sensory details so accessibility teams have a stronger foundation.
The next important evidence will come from participants: whether recognition works under real gallery conditions, whether the route supports independent movement and whether visitors feel the content represents the works meaningfully. If the pilot is assessed transparently, its value could extend far beyond one device or one museum.
ATIM approaches artificial intelligence with curiosity, independence, and balance. We explore its extraordinary potential to empower artists while examining the questions it raises around authorship, copyright, consent, creative identity, transparency, and human expression.
We are neither advocates for AI nor opponents of it. Our focus is how the technology is used—and what that means for artists.
Above all, our commitment is to artists. We look at every development through the lens of what can support, protect, and empower the creative community—identifying opportunities, examining potential risks, and providing artists with the information they need to make informed decisions about their work, their rights, and their future.