Facial Recognition software: Perception vs Reality
Bottom Line: I’ll start by asking the question. Do you trust facial recognition software? If you don’t that makes you normal. On back of San Francisco's recent move to ban the official use of it in the city, there’s been added scrutiny and conversation about its use across the country. I already mentioned that most of us don’t trust the technology based on our current perceptions of it. This despite the continued use of it in consumer technology, banking and cyber security. There’s definitely a disconnect in here somewhere. So where is it? First our perception...
PCMag’s study of the topic showed that only 25% of us trust the technology generally. The area of highest trust...
- Use by law enforcement: 37%
And what we trust least with facial recognition technology...
- Bill pay / financial transactions: 19%
I’m not really sure how to analyze that one. Though it’s probably an inappropriate joke, for the 18% who don’t trust the technology for financial transactions but do for law enforcement – it might literally be a case of he looks guilty. Anyway. Here’s the fact of the matter. The technology is highly accurate, probably more than we give it credit for, but it’s not perfect. Based on government records from the National Institute of Standards and Technology, the accuracy rate is 90%.
And here’s what that 90% means. It doesn’t mean that 1 in 10 people can trick the facial recognition software. It means that out of a national database, a false positive ID would only be incorrect one in ten times.
Are you surprised? Not surprised? Does it change your opinion of the use of technology? We’re all entitled to our opinions, but I thought it’d be helpful to have the facts.