From Greece, to Mexico, to Kenya, to Palestine, borders across the globe have become hotbeds of unregulated technological experimentation, where entire ecosystems of automated “migration management” technologies are being deployed to intensely surveil people on the move with little accountability or oversight. And the dragnet they’ve created is having a devastating human cost.
Instead of being able to exercise their internationally recognised human right to migrate, the vast array of surveillance technologies now deployed against people on the move – including drones, sound canons, robo-dogs, surveillance towers, predictive analytics, biometric data harvesting, lie detectors, heat sensors, high-tech refugee camps, and more, many of which have now been infused with artificial intelligence (AI) – means they are being forced into increasingly desperate and life-threatening situations.
As a result, entire border-crossing regions have been transformed into literal graveyards, while people are resorting to burning off their fingertips to avoid invasive biometric surveillance; hiding in dangerous terrain to evade pushbacks or being placed in refugee camps with dire living conditions; and living homeless because algorithms shielded from public scrutiny are refusing them immigration status in the countries they’ve sought safety in.
In her book, The walls have eyes: Surviving migration in the age of artificial intelligence, refugee lawyer Petra Molnar documents and centres on countless first-hand tales of people facing fear, violence, torture and death at the hands of state border authorities.
“Borders are both real and artificial. They are what historian Sheila McManus calls an ‘accumulation of terrible ideas’, created through colonialism, imperial fantasies, apartheid, and the daily practice of exclusion,” writes Molnar. “The walls have eyes offers a global story of the sharpening of borders through technological experiments, while also introducing strategies of togetherness across physical and ideological borders.
“It is an invitation to simultaneously bear witness to violent realities and imagine a different world – because a new world is possible and ‘hope is a discipline’. We can change the way we think about borderlands and the people who are caught at their sharpest edges.”
Speaking with Computer Weekly, Molnar describes how lethal border situations are enabled by a mixture of increasingly hostile anti-immigrant politics and sophisticated surveillance technologies, which combine to create a deadly feedback loop for those simply seeking a better life.
She also discusses the “inherently racist and discriminatory” nature of borders, and how the technologies deployed in border spaces are extremely difficult, if not impossible, to divorce from the underlying logic of exclusion that defines them.
‘We are Black and the border guards hate us. Their computers hate us too’
For Molnar, technology provides a window into how power operates in society, and whose priorities take precedence.
“So much of the tech absolutely could be used for other purposes,” says Molnar, noting how drones could be used for maritime rescue, while AI could be used to audit immigration decisions or help identify racist border guards.
“Instead, it’s always weaponised and positioned as a tool to oppress an already marginalised group of people. So much of it is about these broader logics of what the migration control system is supposed to be doing – which is preventing people who are unwanted or an ‘other’, or seen as a threat or a fraud, and keeping them away as much as possible, and also to deter people from coming.”
In criminal law, you’re innocent until proven otherwise. But you’re not a refugee unless proven otherwise
Petra Molnar, human rights lawyer and author
Given the socio-technical nature of technology – whereby the technical components of a given system are informed by social processes and vice versa – Molnar says the conceptualisation of borders as a bulwark against “the other” affects how technology is developed for use in border spaces.
This dynamic is summed up by a quote Molnar uses in the book from Addisu – a person on the move, originally from Ethiopia, who has been trying to reach the UK since arriving in Europe two years ago: “We are Black and the border guards hate us. Their computers hate us too.”
In essence, the exclusionary impulse of borders means migration is ultimately framed as a problem: “It’s seen as something to solve, and now we have tech to solve the problem.”
Molnar adds the framing of migration as a problem means states are then able to derogate from their rights because they’re seen as threats or frauds, and are not considered refugees until proven otherwise.
“If you look at criminal law, in most jurisdictions at least, you’re innocent until proven otherwise. But you’re not a refugee unless proven otherwise. We call it the reverse onus principle, where it’s on the person to prove that they are telling the truth, that they have a legitimate right to protection.”
She says if this is the political starting point, “where you assume that everybody is unwelcome unless proven otherwise, and you don’t have a lot of law, and you’re obsessed with technology, it creates this perfect environment for really high-risk tech with pretty much zero accountability”.
Reduced to a data point
To Molnar, the use of surveillance technologies to manage people’s movement across borders is inherently dehumanising. “That’s something I saw as a trend in so many of the conversations I had with people on the move, who were reflecting on being reduced to a data point, or an eye scan or a fingerprint, and already feeling more dehumanised in a system that’s primed to see them as subhuman,” she says.
She adds that automating or even just augmenting migration-related decision-making with algorithms or AI also works to divorce people on the move from their humanity in the eyes of those ultimately making the decisions.
“Instead of looking somebody in the eye, they’re looking at an image of a person or a data point that is, again, divorced from a person’s humanity and the complexity of people’s stories and legal cases.”
Because border spaces are already so opaque, discretionary and characterised by huge power differentials between border officials and people crossing, Molnar says the use of various technologies only makes it more difficult to introduce accountability and responsibility, not just in terms of governance, but on the human level of how it divorces the people carrying out the violence from their own humanity as well.
“When the violence happens way over there as a result of tech, as a result of surveillance, it’s not so immediate, and then maybe not so viscerally felt, even by the decision-makers who are there. That is also a violent practice, this disavowal of responsibility.”
“Technology forces us to not sit in the beautiful complexity of what it means to be a human being, but rather be categorised as a data point in these rigid categories that don’t map onto the messiness of human reality”
Petra Molnar, human rights lawyer and author
This effect is also exacerbated by AI, which has particularly insidious effects in migration-border management contexts because of the way it sees people on the move through the gaze of past prejudices, and essentially projects existing inequalities, biases and power imbalances into the future while treating those discrepancies as an objective truth.
“Ultimately, it’s about putting people in boxes and concretising their experience based on really rigid categories,” she says. “Technology forces us to not sit in the beautiful complexity of what it means to be a human being, but rather be categorised as a data point in these rigid categories that don’t map onto the messiness of human reality.”
By reducing people to rigid categories and classifications, Molnar says it becomes easier to treat them with cold, computerised contempt.
However, she adds, while a logic of deterrence is clearly baked into the global border system, in practice, the use of increasingly sophisticated surveillance systems only works to push people towards increasingly dangerous routes, rather than deter them completely.
“You see that a lot with the border surveillance infrastructure that’s grown up around the Mediterranean and Aegean, but also the US-Mexico corridor. They’ll say, ‘If we introduce more surveillance, then people are going to stop coming’, except that doesn’t work,” she says, adding that people exercising their internationally protected right to asylum are being forced into taking more dangerous routes to reach safe destinations.
“That’s why you see so many people drowning in the Mediterranean or Aegean Sea, or why you have deaths nearly tripling at the US-Mexico border.”
‘Humane’ dehumanisation
Molnar says that while border technologies have a clearly dehumanising effect on people on the move, official justifications for using the tech largely revolve around making the migration process more humane.
“The Democrats [in the US] have become very good at this because they say, ‘Smart borders are more humane, they’re better than the Trump wall and putting babies in cages’. But then, when you start picking it apart, you realise that these policies are also hurting people. Again, the near tripling of deaths at the US-Mexico border since the introduction of the smart wall system, that’s quite telling,” she says, adding that technology often works to obfuscate the extent and seriousness of border violence, hiding under the guise of somehow being more humane: “That’s why it’s important to interrogate the power of the technology.”
In many cases, the deployment of surveillance technologies in migration contexts is not only posited as more humane, but is explicitly justified under the pretext of providing humanitarian support to underdeveloped countries.
“Europe and the US are so implicated in supporting regimes that are very problematic under the guise of humanitarian support. But oftentimes, it’s for border externalisation – it’s to get other actors to do the dirty work for you,” she says, adding that the European Union (EU), for example, regularly provides funding and tech to various paramilitaries on the African continent that are involved in border enforcement, as well as coastguards and border force teams in countries like Libya and Niger.
“If the frontier is moving further and further away, it makes it easier for ‘Fortress Europe’ to remain unassailed.”
Molnar adds that these kinds of humanitarian justifications for increasing border tech deployments are also being pushed by the third sector and non-governmental organisations.
Although people working in these organisations are often well-intentioned, Molnar says international organisations such as the United Nations (UN), Unicef or the World Food Programme have “huge normative power” over the idea that “more data is better”, and are therefore a massive driving force behind normalising a lot of the border tech currently in use.
“The refugee camps in Kenya, like Dadaab and Kakuma, were some of the first places that had biometric registration. If you look at the Global Compact on Migration, which is this big international document that was put together a few years ago, the first point is ‘more data’. That’s quite telling,” she says.
“When you see, for example, what the United Nations High Commissioner for Refugees did with the Rohingya refugees – they collected so much data, and then inadvertently shared it with the Myanmar government, the very government that the refugees are trying to flee from.
“But how did that happen? I think we need to query what happens in this ‘third space’ of international actors too, not just states or the private sector.”
Surveillance pantomime
In her book, Molnar notes that while the various technologies of surveillance and control deployed at borders work well because of their diffuse, omnipresent nature, they often don’t even have to be that effective in achieving the goals of state authorities, as “their spectre and spectacle changes our behaviour, modifies our thinking, and adds to a general sense of unease, of always being watched”.
Highlighting her visits to the Evros region between Greece and Turkey, Molnar says it’s not always clear what the tech is doing, especially when it comes to some of the more obscure AI-driven tools being used.
“It’s almost like it’s the ‘performance’ of the tech that’s more important,” she says, adding that while it is certainly true that the tech does directly affect negative outcomes for people on the move, it’s unclear if this is because the tech is doing its job, or because of the powerful effect “security theatre” has on people’s behaviour.
“The performance of surveillance and securitisation is what is quite powerful. But ultimately, I think so much of it is about politics … you feel the power of the surveillance, even if it’s not really there. You feel the paranoia.”
Molnar adds that this dynamic is further enabled by the legislative and governance frameworks around border technologies, which essentially work to shield the states and corporations involved from any meaningful accountability, as invoking the spectre of “national security” allows them to shut down any scrutiny of the tech they’re deploying.
While the governance of border technologies globally is already characterised by extreme opacity, this gets even worse “as soon as the national security paradigm is invoked, because then there are even fewer responsibilities that a state has, for example, to citizens or concerned researchers to tell them what’s actually happening”.
Molnar adds that while governance and regulation can help improve transparency within a system that is designed to be opaque, there have been “disappointing trends” in recent years.
“I think a lot of us – perhaps naively – are hoping that the European Union’s AI Act might be a strong force for good when it comes to putting up some guardrails around border tech in particular,” she says, adding that while it contains positive measures on the face of it – including a risk matrix and allowing for certain technologies to be coded as high risk – the national security carve-outs mean these responsibilities placed on various actors by the legislation simply do not apply in border spaces.
“As soon as you can say something’s national security, so the law doesn’t apply in the same way, what good is a piece of legislation like that?”
Artificial and colonial borders
Highly critical of the way artificial borders are seen as natural phenomena when they are, in fact, historically novel social constructs, Molnar says the surveillance apparatus being deployed worldwide helps to maintain and reinforce imperialist power dynamics.
“So many people think that borders and bordering as a practice is something that’s always been with us when, in fact, it’s a social construct, and borders have been shifting and changing since time immemorial,” she says. “There were times where borders really weren’t a thing, where people could just travel freely from place to place, and the current reality of hard border control is actually a very recent phenomenon.”
Relaying a recent trip to Naco, a small border community in Arizona, Molnar says she was struck by how porous the US-Mexico border was even just a few decades ago, with locals sharing tales of how they are no longer able to move across town since it was bifurcated by a huge wall.
“People would be playing volleyball across the border, and now there’s this hulking piece of infrastructure. It seems so intractable like it was always there, but that’s not the case at all.”
For Molnar, the horrific human impacts of border technologies therefore ultimately run along and reinforce “colonial delineations” with who is seen as a “worthy” immigrant and who becomes the ultimate “persona non grata” being “mapped along the lines of Western imperialism, white supremacy and apartheid”.
She warns that these modes of thinking mean when it comes to the development of border or migration management technologies, people on the move and their needs will always be thought of last.
“Why can’t they just use AI to help us with all the forms?” she says, only half joking. “Instead, it’s visa triaging, and robo-dogs, and AI-powered lie detectors. Why don’t you just talk to people on the move? What do they need?
“But again, it’s all about power, and it’s power brokers that are the ones making decisions. It’s the states, it’s the private sector, it’s the UN and international organisations. It’s definitely not affected communities.”
Further highlighting the example of AI-powered lie detectors in airports – which were developed by the EU’s Horizon 2020 project – Molnar says the academics involved (who she spoke with directly) did not take into account the way people on the move may act differently due to trauma, which affects their memory and how they tell stories or relay information, or cultural differences in the ways they communicate.
“I remember talking to this group of academics, and they were so distressed. They were like, ‘We didn’t think about any of this’, and I said, ‘How could you not? Did you not talk to a single refugee lawyer or a refugee before designing this?’ That’s disturbing to me,” she says.
“Affected communities are the last possible kind of stakeholder in this conversation, and I think we need to flip that completely.”
Optimism over despair
It is an open question for Molnar whether it is even possible to introduce new technologies into border spaces that don’t support their inherently “violent ideals of exclusion”, noting that while it is entirely possible to imagine genuinely helpful uses of tech, a confluence of powerful interests is preventing this from happening.
Why can’t they just use AI to help with all the forms? Instead, it’s visa triaging, and robo-dogs, and AI-powered lie detectors
Petra Molnar, human rights lawyer and author
“So much of the money in the ‘border industrial complex’ that’s grown up around border tech is there to support states and the private sector in their goal to keep people out, rather than using even a fraction of this money to either make the system better as it is right now, or even support tech development for communities, by communities.”
However, Molnar says there are actions that people can and are taking to help support people on the move facing violence at borders.
She adds while there are already calls for stronger regulation of border technologies to hold governments accountable, and civil society and journalists have a role in asking difficult questions about their tech deployments, one option could be to take a smaller, more localised approach.
Highlighting various municipalities in the US that have banned facial recognition technology, Molnar says a similar “community approach” could be taken with regards to the various technologies being deployed in border spaces.
Despite this, she notes “that also isn’t enough” as any approach would also have to look at the problem holistically, given border tech specifically operates at both a national and international scale, creating a tension and disconnect between people on the ground and those holding the political levers of power.
Given the rapid environmental degradation taking place worldwide, Molnar adds the use of technology to push people away and strengthen borders simply will not stop people from migrating.
An alternative approach is therefore urgent, which Molnar says must include challenging the existing laws (while recognising the clear limits of our current legal systems); co-opting and co-designing the tech in the interests of people on the move, rather than corporations and state authorities; and creating participatory institutions self-directed by people on the move (based on the principle ‘nothing about us without us’).
“I think so much of it comes down to seeing one another as fully human, and leading from a place of curiosity rather than fear.”
Molnar concludes that while the situation at borders around the globe may be bleak, “there are always people who make choices to show up”.
Whether that be people launching their own search and rescue boats in the Mediterranean, going into the Sonoran Desert to do “water drops” for those crossing the dangerous terrain, or farmers sheltering people on the move in the corridor between Poland and Belarus – all of which pose a real threat of arrest and loss of liberty – Molnar says it’s ultimately about human-to-human interaction and finding ways of moving past differences: “There is always a choice.”