museum visitors are being monitored by AI-powered cameras

New digital art experiences such as Outernet in London have been making headlines recently with their large attendance figures compared with more traditional museums; last year, for example, Outernet reported having almost 6.3 million visitors compared with the British Museum’s 5.8 million.

But how does Outernet count visitors wandering into its open-sided building on a busy junction in central London? It uses video surveillance. Or, more accurately, it employs the video analytics provider BriefCam to track visitors using deep learning, an artificial intelligence (AI) method that teaches computers to process data like the human brain.

A clip on the company website asks: “What if you could drive exponential value from your video content by making it searchable, actionable, and quantifiable?” as a man behind a computer zooms in on freeze frames of faces and car number plates. Forensic investigations, vehicle make and model recognition, and face recognition are among the services offered by BriefCam, alongside crowd management and “elevated guest engagement”.

“We are able to track gender and are exploring even more sophisticated methodologies around dwell time and how this is influenced by what the public come to see, time of day, weather, and cultural calendar moments,” says an Outernet representative. “We are proud to be working towards offering some of the most advanced campaign reporting in the world. This, in addition to our neuroscience research launched last year, as well as our more traditional research methodologies, including on-site qualitative and quantitative audience measurement, provides us with robust insights about visitors and their experiences.” They add that all data recorded by Outernet is “anonymised and of course protected”.

BriefCam told The Art Newspaper that it draws the line at tracking race, which is in line with the UK General Data Protection Regulation’s (GDPR) ban on the “processing of personal data revealing racial or ethnic origin”. The company added that “it is BriefCam’s expectation that users apply technology in accordance with applicable laws and while upholding civil rights. We are committed to protecting privacy, preventing abuses, and the ethical and responsible use of surveillance technologies.”

The private Glenstone museum, in Potomac, Maryland, uses BriefCam to beef up security, to implement “smarter workflows for better visitor experiences”, BriefCam explains on its website, and to track real-time occupancy in its galleries and grounds. If someone gets too close to a work of art or strays from the outdoor trails, the system will send an automatic alert so Glenstone can “respond to the behaviour in order to preserve property, safety, and ambiance”.

Some museums, however, are only concerned with employing technology to count visitors accurately. Back in 2018, The Art Newspaper revealed that faulty equipment provided by an external company for the National Portrait Gallery (NPG) in London failed to register more than 600,000 visitors over 12 months. The museum’s current people counter, FootfallCam, is seemingly working. A spokesperson for the NPG says it relies on “a well-established programme of visitor research which provides insights into visitor experience”, and has no plans to implement FootfallCam’s other data analytics services, such as facial recognition sensors that detect “sentiment”: in other words, cameras that discern the emotion etched on someone’s face upon leaving an exhibition space to gauge visitor satisfaction.

A spokesperson for Tate says the institution only uses cameras to count visitors and gathers “more detailed insights and demographic information through visitor research and surveys”. The Royal Academy of Arts (RA) in London uses another people tracker called V-Count to capture footfall data in the museum’s non-ticketed spaces and to work out things like “the proportion of site visitors that come but don’t buy a ticket for an exhibition”, according to an interview on V-Count’s website with Carlo Palazzi, the RA’s former insights and analytics lead. V-Count is also used by the global immersive exhibition brand Arte Museum in the US, Asia and the Middle East, as well as Australia’s National Gallery of Victoria and Norway’s Munch Museum.

“Our innovative AI-driven technology doesn’t track or label individuals in any way, shape, or form,” says a spokesperson for V-Count when asked how it safeguards recorded information against potential misuse. “Instead, it provides key statistical data on an hourly or daily basis. It’s important to note that this data is completely anonymised, and therefore it is not linked to any specific individual. This ensures our technology cannot be used for discriminatory or harmful purposes.”

The company does not disclose the locations of its servers but stresses its adherence to GDPR standards to ensure a high level of data protection for users. “Our advanced sensors have been developed to determine both gender and age without the need to capture, record, store or transfer any video or image data so that no personal data is ever captured or compromised,” V-Count’s spokesperson adds.