Trustworthy AI Best Practices: Lessons Learned from the Rite Aid Facial Recognition Ban [AI Today Podcast]

Facial recognition falls under the recognition pattern of AI. On it's own the technology is neutral. However, it's the application and use (or misuse) of this technology that can have severe consequences. Rite Aid is one of the most recent companies to come under fire for their use of facial recognition. In this episode of the AI Today podcast we explore what happened that that caused the FTC to give retailer Rite Aid fa facial recognition ban for 5 years. Why is facial recognition being banned? Between 2012-2020 Rite Aid was using facial recognition technology in it's stores to help identify customers who may have been engaged in shoplifting or other problematic behavior. Problem was, the system was falsely identifying consumers, particularly women and people of color, as shoplifters. So, what went wrong? The facial recognition system Rite Aid used clearly didn’t have enough diverse training data and enough good quality data/images. The images were of such low quality they framed innocent people. Additionally, the systems lacked sufficient human oversight with employees not being properly trained and blindly trusting the AI system. Also, there was a clear lack of AI system controls and Trustworthy AI framework in place. How accurate is facial recognition? Facial recognition technology, like all AI technology, is probabilistic. AI systems are probabilistic as they provide predictions with differing levels of probability and are best suited to probabilistic problems where data can be highly variable. So, this means that facial recognition systems will never be 100% accurate. This means that AI systems need constant human supervision. Never blindly trust an AI system, especially one that can cause personal harm. Don’t drive over the cliff because the AI system tells you to. How do I make AI more Trustworthy? Buying AI solutions from a vendor doesn’t absolve you of responsibility to make sure the technology is being used properly. The core of the complaint against Rite Aid was “that the company failed to take reasonable measures to prevent harm to consumers”. It doesn’t matter if you bought the AI system from a vendor or built it yourself. Your AI system, your problem. What is a Trustworthy AI Strategy? In light of this news you may be asking yourself: Can I opt out of facial recognition? Can you block facial recognition? Tune into this episode to learn more about what happened and what this FTC facial recognition ban ruling mean for you. Snow Notes: Cognilytica newsletter Rite Aid Banned from Using AI Facial Recognition After FTC Says Retailer Deployed Technology without Reasonable Safeguard FREE Intro to CPMAI mini course Seven Patterns of AI AI Today Podcast: Trustworthy AI Series: The Layers of Trustworthy AI Trustworthy AI Framework Training & Certification