The UK government has turned a US-based startup specialized in artificial intelligence as part of its pledge to stop small-boat crossings. Experts have already pointed out the legal and logistical challenges of the plan.
In a new effort to address the high number of Channel crossings, the UK Home Office is working with the US defense startup Anduril, specialized in the use of artificial intelligence (AI).
A surveillance tower has already been installed at Dover, and other technologies might be rolled out with the onset of warmer temperatures and renewed attempts by migrants to reach the UK. Some experts already point out the risks and practical loopholes involved in using AI to identify migrants.
"This is obviously the next step of the illegal migration bill," said Olivier Cahn, a researcher specialized in penal law.
"The goal is to retrieve images that were taken at sea and use AI to show they entered UK territory illegally even if people vanish into thin air upon arrival in the UK."
The "illegal migration bill" was passed by the UK last month barring anyone from entering the country irregularly from filing an asylum claim and imposing a "legal duty" to remove them to a third country.
Who is behind Anduril?
Founded in 2017 by its CEO Palmer Luckey, Anduril is backed by Peter Thiel, a Silicon Valley investor and supporter of Donald Trump. The company has supplied autonomous surveillance technology to the US Department of Defense (DOD) to detect and track migrants trying to cross the US-Mexico border.
In 2021, the UK Ministry of Defence awarded Anduril with a £3.8-million contract to trial an advanced base defence system. Anduril eventually opened a branch in London where it states its mission: "combining the latest in artificial intelligence with commercial-of-the-shelf sensor technology (EO, IR, Radar, Lidar, UGS, sUAS) to enhance national security through automated detection, identification and tracking of objects of interest."
According to Cahn, the advantage of Brexit is that the UK government is no longer required to submit to the General Data Protection Regulation (RGPDP), a component of data protection that also addresses the transfer of personal data outside the EU and EEA areas.
"Even so, the UK has data protection laws of its own which the government cannot breach. Where will the servers with the incoming data be kept? What are the rights of appeal for UK citizens whose data is being processed by the servers?", he asked.
Also read: Under French coastguard's eye, migrants cross into English waters

'Smugglers will provide migrants with balaclavas for an extra 15 euros'
Cahn also pointed out the technical difficulties of identifying migrants at sea. "The weather conditions are often not ideal, and many small-boat crossings happen at night. How will facial recognition technology operate in this context?"
The ability of migrants and smugglers to adapt is yet another factor. "People are going to cover their faces, and anyone would think the smugglers will respond by providing migrants with balaclavas for an extra 15 euros."
If the UK has solicited the services of a US startup to detect and identify migrants, the reason may lie in AI’s principle of self-learning. "A machine accumulates data and recognizes what it has already seen. The US is a country with a significantly more racially and ethnically diverse population than the UK. Its artificial intelligence might contain data from populations which are more ethnically comparable to the populations that are crossing the Channel, like Somalia for example, thus facilitating the process of facial recognition."
People are going to be identified and there are going to be errors. If a file exists, there needs to be the possibility for individuals to appear before justice and have access to a judge.
For Cahn, it is not capturing the images which will be the most difficult but the legal challenges that will arise out of their usage. "People are going to be identified and there are going to be errors. If a file exists, there needs to be the possibility for individuals to appear before justice and have access to a judge."
Also read: UK government could be considering using barges to house migrants offshore
A societal uproar
In a research paper titled “Refugee protection in the artificial intelligence Era”, Chatham House notes "the most common ethical and legal challenges associated with the use of AI in asylum and related border and immigration systems involve issues of opacity and unpredictability, the potential for bias and unlawful discrimination, and how such factors affect the ability of individuals to obtain a remedy in the event of erroneous or unfair decisions."
For Cahn, the UK government’s usage of AI can only be used to justify and reinforce its hardline position against migrants. “For a government that doesn't respect the Geneva Convention [whose core principle is non-refoulement, editor’s note] and which passed an illegal migration law, it is out of the question that migrants have entered the territory legally.”
Identifying migrants crossing the Channel is not going to be the hardest part for the UK government. Cahn imagines a societal backlash with, "the Supreme Court of the United Kingdom being solicited, refugees seeking remedies to legal decisions through lawyers and associations attacking".
He added there would be due process concerning the storage of the data, with judges issuing disclosure orders. “There is going to be a whole series of questions which the government will have to elucidate. The rights of refugees are often used as a laboratory. If these technologies are 'successful', they will soon be applied to the rest of the population."
Also read: UK spending far more on migrants than on aid for Africa, official figures reveal