Charity say using AI on refugees "not the way to go"
A new system will be trialled in Dover that uses AI to determine ages of asylum seekers.
A refugee charity have said they don't think AI is the right way to go, following government plans to include AI into UK border security from 2027.
Last week the Home Office announced their plans to use AI to check ages, with a post on Facebook Stating: 'AI technology will be rolled out at UK borders to prevent illegal migrants from falsely claiming to be under 18.'
'Currently, false age claims can be used to exploit the system and delay detention and removals, while diverting vital safeguarding support away from genuine asylum-seeking children.'
'In the year ending March 2026, over 6,400 migrants claiming to be children were age assessed at the border, with 43% found to be adults. '
'Strengthened checks will ensure adults who attempt to cheat the system and do not have a legitimate right to be in the UK are identified, detained and removed without delay.'
'Genuine children will be given the protection they need'
But speaking to Greatest Hits Radio Kent, CEO of Care4Calais Steve Smith said:
"The theory is that by using an automated process, by using AI, it will have certain advantages over the current system.
"The current system, we know is inadequate. We know a huge number of mistakes are made.
"And clearly that's really important because if you get it wrong, what you're effectively doing is sending a child into an adult situation into, for example, a hotel occupied entirely by other adults.
"So AI hopefully would get around that. It would hopefully speed things up.
"It would hopefully be an objective rather than a subjective system.
"But there are flaws in that, it is by no means perfect and we as a charity consider that it's not the way to go."
He added that AI would be basing it's conclusions solely on the appearance of the individual, which may not be enough data.
"We know that AI works off the information it's been given.
"And we know that when it's got that information, it's largely based on Western demographics.
"We've seen this when there's consideration about AI being used for facial recognition, potentially for criminal activity.
"If you're coming from Somalia, if you're coming from Afghanistan, Iran, Iraq, the facial recognition is not part of the background library of what this AI is working on.
"So it's obviously a chance for making mistakes. Added to which, these individuals have often been through extreme trauma.
"They may have faced near starvation. Their faces may be somewhat weathered by sleeping rough maybe on the coast of northern France for months at a time. They've encountered danger. They may have been in detention in their country of origin.
"They may have been in detention in their country of origin. They may have been tortured.
"That itself has an effect on the physical characteristics of the face. And AI won't have been schooled in working out how to do that."
He added that there was also a worry about AI making a wrong decision, and not being challenged the same way a person would be by their peers.