Discover more from Last Week in AI
Last Week in AI News #76
Hopefully, you’re already wearing a face mask when you go outdoors. But just in case you needed another reason to keep that mask on, it just might prevent you from being identified by a facial recognition system. Facial recognition systems are known to be faulty, due to the quality of the products and issues like bias. They can have worryingly high error rates, creating the potential for wrongful arrests and other mishaps. But those error rates may work in your favor as well: a study by the US National Institute of Standards and Technology (NIST) found that wearing a mask that covers the nose and mouth can cause the error rate of widely used facial recognition algorithms to spike to between 5 and 50 percent. Mei Ngan, an author of the report, did note that the researchers focused on algorithms developed before the pandemic, and later plan to test the accuracy of algorithms developed with mask wearing in mind.
American citizens have already been alerted to the concerning uses of facial recognition systems by law enforcement and various other groups, with services most notably provided by companies such as Clearview AI. As it turns out, the use of facial recognition might be even more pervasive than we realized: a Reuters investigation found that over eight years, American drugstore chain Rite aid Corp “quietly added facial recognition systems to 200 stores across the United States, in one of the largest rollouts of such technology among retailers in the country.” Reuters found that the technology was deployed in mostly lower-income, non-white neighborhoods in New York City and Los Angeles. Furthermore, for over a year, “the retailer used state-of-the-art facial recognition technology from” DeepCam LLC, a company “with links to China and its authoritarian government.” Indeed, while Rite Aid boasts the most extensive use, it is not the only retailer who has used or tested facial recognition technology–others include Walmart and Home Depot.
As would be expected, issues of AI bias have already bared their ugly face. Tristan Jackson-Stankunas was ordered to leave a Rite Aid store in September 2016 when shopping for an air freshener because a manager claimed he had an image of Jackson-Stankunas stealing from another Rite Aid store. Jackson-Stankunas told Reuters he was not the man in the image, and their only similarity was that they were both African American. It will be unsurprising to see wrongful arrests and other adverse consequences in the future if such technology is deployed in retail without immense caution.
Advances & Business
Startup and Academics Find Path to Powerful Analog AI - Engineers have been chasing a form of AI that could drastically lower the energy required to do typical AI things like recognize words and images
Instrumental raises $20M to scale its AI-powered manufacturing tech - This morning Instrumental, a startup that uses vision-powered AI to detect manufacturing anomalies, announced that it has closed a $20 million Series B led by Canaan Partners. The company had previously raised $10.3 million across two rounds, including a $7.5 million Series A in mid-2017.
How India is using Artificial Intelligence to combat COVID-19 - Artificial Intelligence (AI) has been one of the biggest technology success stories of the past decade. As the COVID-19 pandemic spread across the globe, researchers and entrepreneurs stepped up to devise new ways to combat it.
Artificial intelligence could speed up and improve Alzheimer’s diagnosis - Artificial intelligence (AI) could help to diagnose Alzheimer’s faster and improve patient prognosis, a new study from the University of Sheffield has revealed.
AI model developed to identify individual birds without tagging - For even the most sharp-eyed of ornithologists, one great tit can look much like another. But now researchers have built the first artificial intelligence tool capable of identifying individual small birds.
Concerns & Hype
Service that uses AI to identify gender based on names looks incredibly biased - Some tech companies make a splash when they launch, others seem to bellyflop. Genderify, a new service that promises to identify someone’s gender by analyzing their name, email address, or username with the help AI, looks firmly to be in the latter camp.
AI-Powered “Genderify” Platform Shut Down After Bias-Based Backlash - Just hours after making waves and triggering a backlash on social media, Genderify - an AI-powered tool designed to identify a person’s gender by analyzing their name, username or email address - has been completely shut down.
The problems AI has today go back centuries - Algorithmic discrimination and “ghost work” didn’t appear by accident. Understanding their long, troubling history is the first step toward fixing them.
Expert Opinions & Discussion within the field
The field of natural language processing is chasing the wrong goal - Researchers are too focused on whether AI systems can ace tests of dubious value. They should be testing whether systems grasp how the world works.
Where Is Robotics Heading? Perspectives From iRobot (Colin Angle), Stanley Black & Decker, And Robots In Service Of The Environment - The dream of robots and intelligent machines that can perform a wide array of tasks has been around in the common visions and fantasies of people for centuries. Machines that can do the work of people without having the failings of people is one of those long-sought visions of the future.
Philosophers On GPT-3 - Nine philosophers explore the various issues and questions raised by the newly released language model, GPT-3, in this edition of Philosophers On, guest edited by Annette Zimmerman.