When Facial Recognition Tech Is Wrong
Uncanny Valley | WIRED18 Maalis 2022

When Facial Recognition Tech Is Wrong

Like a lot of tech solutions to complex problems, facial recognition algorithms aren't perfect. But when the technology is used to identify suspects in criminal cases, those flaws in the system can have catastrophic, life-changing consequences. People can get wrongly identified, arrested, and convicted, often without ever being told they were ID’d by a computer. It’s especially troubling when you consider false identifications disproportionately affect women, young people, and people with dark skin—basically everyone other than white men.

This week on Gadget Lab, WIRED senior writer Khari Johnson joins us to talk about the limits of facial recognition tech, and what happens to the people who get misidentified.

Show Notes:

Read Khari’s stories about how facial recognition tech has led to wrongful arrests that derailed people’s lives. Here’s Lauren’s story about Garmin’s Fenix smartwatch. (And here’s WIRED’s review of the latest model.) Arielle’s story about the wave of shows about Silicon Valley tech founders is here.

Recommendations:

Khari recommends hoagies. Lauren recommends Garmin smartwatches. Mike recommends the show The Dropout on Hulu.

Khari Johnson can be found on Twitter @kharijohnson. Lauren Goode is @LaurenGoode. Michael Calore is @snackfight. Bling the main hotline at @GadgetLab. The show is produced by Boone Ashworth (@booneashworth). Our theme music is by Solar Keys.

Learn about your ad choices: dovetail.prx.org/ad-choices

Jaksot(426)

BIG INTV: The Onion's Ben Collins On How The Truth Can Pay Dividends

BIG INTV: The Onion's Ben Collins On How The Truth Can Pay Dividends

A year after relaunching The Onion in print, CEO Ben Collins sits down with Katie to talk about why “going into something and not ruining it is bravery.” He tells her the first order of business to ge...

2 Syys 202537min

WIRED Roundup: Meta’s AI Brain Drain

WIRED Roundup: Meta’s AI Brain Drain

In today’s episode, Zöe is joined by WIRED’s Leah Feiger to run through five of the best stories you need to know about — from how AI is eliminating entry level jobs to how a secretive Democrat group ...

29 Elo 202512min

This Is How You Log Off

This Is How You Log Off

As the summer comes to an end, we’re thinking about what unplugging, going offline — the ultimate luxury — looks like in 2025. With a growing slate of apps aiming to reduce our screen time and some te...

28 Elo 202525min

BIG INTV: Alexis Ohanian Says We Can Build A Less Toxic Internet

BIG INTV: Alexis Ohanian Says We Can Build A Less Toxic Internet

The Reddit co-founder sits down for a wide-ranging conversation, from buying his first video game to meeting his future co-founder on the first day of college.  When he’s not investing in the future o...

26 Elo 202548min

WIRED Roundup: Chip Manufacturers Bonanza

WIRED Roundup: Chip Manufacturers Bonanza

In today’s episode, Zöe is joined by WIRED’s Louise Matsakis to run through five of the best stories we published this week — from OpenAI’s being poised to become the highest valued startup in history...

22 Elo 202519min

How To Become a Vibe Coder

How To Become a Vibe Coder

Vibe coding is everywhere, and it’s already drastically changing the tech industry — from shaping how software gets made to who gets hired. So back in July, our very own Lauren Goode went on a journey...

21 Elo 202529min

Introducing: The BIG INTERVIEW with WIRED's Katie Drummond

Introducing: The BIG INTERVIEW with WIRED's Katie Drummond

A conversation series that only WIRED could host, led by WIRED’s Global Editorial Director, Katie Drummond. Every week, Katie sits down with influential figures in culture, politics, business, science...

19 Elo 20252min

WIRED Roundup: Why GPT-5 Flopped

WIRED Roundup: Why GPT-5 Flopped

In today’s episode, Zöe is joined by WIRED’s Jake Lahut to run through five of the best stories we published this week — from how the Trump administration is creating and sharing memes to make fun of ...

15 Elo 202517min