9.21.20 – SIW –
them to remove bias on the part of those monitoring the footage.
Photo: Percepta
Company makes $1M investment in Percepta Labs “ethical AI” technology
ADT Commercial is investing in Percepta Labs, an artificial-intelligence technology startup out of Philadelphia. The $1M investment will fund the commercialization and application of the startup’s cutting-edge AI technology to help detect and deter shoplifting. According to a press release, ADT Commercial will work directly with Percepta to further test and develop the application of Percepta’s proprietary, socially responsible AI technology in commercial environments and facilitate the launch when the solution is market-ready.
Founded by Philippe Sawaya, Neil Gramopadhye, and Jonathan Mak while pursuing their undergraduate degrees at the University of Pennsylvania, Percepta’s proprietary “ethical AI” technology responds to emerging concerns surrounding the use of facial recognition technologies and is designed to completely anonymize individuals’ demographic features – including race, gender, and age. People appear as three-dimensional “drawings” (see the photo), thus eliminating bias. Using advanced action recognition techniques, the software analyzes shoppers’ movements using existing security camera feeds to detect shoplifting incidents in real-time by scanning for probable shoplifting or precursory shoplifting behavior, and giving the anonymous people a theft probability score that loss prevention professionals can then act upon.
“We pride ourselves on being at the forefront of change in our industry,” ADT EVP Dan Bresingham said in a statement. “We believe Percepta could signal a tidal shift in how we approach AI, facial recognition, and data collection, and we look forward to seeing how their technology continues to develop.”
In addition to taking a socially responsible approach to AI, ADT says Percepta’s technology shows tremendous promise when it comes to loss prevention accuracy. “Early tests by Percepta indicate the technology could be more than two times more accurate than human loss prevention personnel in detecting shoplifting behaviors, translating into improved margins, decreased shrink, and a better experience for customers, while eliminating biases based on race, gender and age,” the release says.