, ,

You might not always understand your computer, but increasingly, it understands you. The specific field of artificial intelligence (AI) that pertains to “computer vision” involves teaching computers how to “see” and interpret the world and its residents, much like humans do. And they’ve gotten very good at it, at least in some ways.

Retailers have outfitted computers with high-definition cameras and put them to work doing things like spotting customers who skip items in the self-checkout line, helping customers try on clothes in a virtual dressing room, and even providing recommendations to customers about things they may want to buy. But of course there are a few catches.

A big one is that the computer vision programs are, it turns out, rife with bias: racial bias, gender bias, and age bias, to name a few. Most facial recognition software, for example, is highly accurate when it comes to identifying people with lighter skin, but not so much for people whose skin is darker. The technology also presents serious privacy concerns.

These problems are so pervasive that Amazon and Microsoft both announced they would stop selling their facial recognition tech for use by law enforcement agencies. IBM went so far as to announce, in a letter to Congress, that it would no longer offer, develop, or even research facial recognition technology because of its problematic uses, noting: “IBM firmly opposes and will not condone uses of any technology, including facial recognition technology offered by other vendors, for mass surveillance, racial profiling, violations of basic human rights and freedoms, or any purpose which is not consistent with our values and Principles of Trust and Transparency.” The company’s CEO Arvind Krishna went on to propose starting a “national dialogue” about this particular technology’s good and bad uses and how to protect against the latter.

So what does all this mean for retailers and customers? It may be too soon to say how precisely the ethical and logistical issues will be resolved (if at all). Yet experts still keep issuing their predictions about the vast growth of the industry in coming years. According to a report that Allied Market Research released in January 2022, the value of the worldwide computer vision market will increase from $9.45 billion in 2020 to $41.11 billion by 2030. Seems like enough time to train the computers to be less racist … we hope?

Discussion Questions

  1. What are some ways retailers can use computer vision to engage and attract more customers, or increase their profits?
  2. What are some of the key problems associated with using computer vision?
  3. Do you feel comfortable with law enforcement and/or retailers using computer vision applications on you?
  4. What should be done to eliminate bias from computer vision applications?

Source: Cleber Ikeda, “How Far Should Retailers Go With the Use of Computer Vision?” RetailWire, February 17, 2022; Chris Ciabarra, “AI vs. Privacy: Does Computer Vision Violate Your Right to Privacy?” Forbes, December 18, 2019; Nick Babich, “What Is Computer Vision & How Does it Work? An Introduction,” Xd Ideas, July 28, 2020; Shemmy Majewski, “5 Ways Computer Vision Is Transforming Retail Industry,” DLabs.AI, October 13, 2021; Jenn Halweil and Skylar Walters, “From Black to White to Technicolor: Debiasing Racism in Computer Vision,” Interesting Engineering, August 9, 2021; Jay Peters, “IBM Will No Longer Offer, Develop, or Research Facial Recognition Technology,” The Verge, June 8, 2020; Allied Market Research, “Computer Vision Market to Reach $41.11 Bn, Globally, by 2030 at 16.0% CAGR: Allied Market Research,” PR Newswire, January 13, 2022