‘AI is making me invisible’: Woman with facial paralysis says 'exclusionary' tech is leaving disabled people behind
“I'm a real person with a real face and I'm still getting rejected because my appearance doesn't look the way a computer expects it to.”
A woman with facial paralysis says AI is making her feel “invisible” as facial-recognition technology fails to recognise her and is locking her out of everyday life.
Listen to this article
Caoimhe Reddy, 29, originally from Cork but now living and working in London, has Bell’s palsy and says she is constantly battling biometric systems like Face ID and e-gates at border control to prove she is a real person.
“Things that work for regular people don’t work for me,” she told LBC.
“My face now moves and looks different to how it looks in a photo so the biometrics systems essentially can't recognise me.
“I have synkinesis and nerve damage so my face doesn't move the way - and I hate to use this word - the way a normal person's does."
She explained her eyes blink at different rates and she has droop in her mouth depending on how severe her symptoms are that day.
Read more: Rogue OpenAI agent hacked Australian government website in world first security breach
‘Dehumanising’
Caoimhe said the repeated failures have left her feeling self-conscious and frustrated.
“Something that's meant to be convenient essentially locks me out, which is dehumanising.
“I was already deeply insecure and deeply self-aware about my condition. I wasn't waiting on a computer to tell me that I didn't look like myself. I already knew that I didn't look like myself, and I was trying to process that."
She explained: “It’s created this sense of shame, which obviously I had nothing to be ashamed of. This is just something that happened to me.
“But I would be walking off to passport control with dread in my stomach.”
Caoimhe said she is worried about the long-term consequences of the increasing use of biometric data to prove identification.
"I have concerns once it becomes infrastructure rather than just convenience, as the failure of what I am currently experiencing has much bigger consequences."
She said: “I'm a real person with a real face and I'm still getting rejected because my appearance doesn't look the way a computer expects it to.”
“My identity hasn't changed. It's literally just the technology's ability to recognise me has," she added.
‘Stuck in a loop’
Caoimhe said attempts to resolve the issue can leave her “stuck in a loop” of emails, phone calls and referrals - and rarely speaking to a real person about it.
“It's crazy to me then that like we've just done away with that human element. Sometimes there's no person on the other side. They've actually done away with customer service representation in favour of an AI representative."
She explained that for a recent right-to-work check, she was asked by a company to scan her passport, then submit an image of herself smiling.
Caoimhe contacted the company after being unable to complete the process.
“I told them I can't pass this. And they said, ‘Oh no, don't worry about it, it's super convenient, just take a selfie and submit it.
“So I replied saying I can't do what it's asking me to, and your system has just assumed that absolutely everyone can take a selfie and smile.
“I can smile, but it doesn't look like what a normal person's smile looks like because I don't have movement on one side of my face.”
Caoimhe said she finds it troubling that people’s faces were being treated as a “static identifier”.
“They change through everything, like illness, scars, ageing, even weight loss."
She has also found tech’s response to her issues disappointing.
“I've yet to get in touch with someone who says here’s an alternative or here's the solution or here's what we're doing.
“Overwhelmingly, I get so sorry about that. We'll take that away. We'll see what we can do, or we'll figure out what we can do or leave this with us so we can figure it out."
She added poignantly: “But I think that the technology doesn't have to necessarily deliberately exclude someone for it to be exclusionary.”
‘At what cost?’
Caoimhe has noticed an increasing use of facial recognition and AI tools over the last year.
“I have definitely noticed, I would say in the last like 12 months, there's just an increasing implementation of things like facial recognition and AI tools in things like passport control, biometric checks becoming a convenience thing.
“All of a sudden that's increasingly being done away with. And I get that it might be more efficient, but at what cost?”
Phyllida Swift, CEO of Face Equality International, has called for "greater inclusion".
"Harms posed by AI and in particular facial recognition technology are something the facial difference community are increasingly telling us about here at Face Equality International," she said.
"We're so glad that people are speaking up about these issues.
"As an international alliance, we're calling for greater inclusion, so that the voices of the FD community are heard in the further development and rollout of AI technology.
"This is because we know that there are real life accessibility issues right now across many different contexts, from airports to job applications to banking and secure apps.
"Together with Northumbria University, we are expanding our Visibly Invisible campaign to enable people to report issues with FRT and to hold companies to account when developing and adopting these technologies."
A government spokesperson said: “We want our digital services to be accessible to everyone, and we recognise how frustrating and upsetting it can be when someone is unable to use facial verification.
“Alternative routes to verify identity are available for anyone experiencing difficulties, and we will continue working to ensure our services are as accessible and inclusive as possible while maintaining the security checks needed to protect the public."