AI in the sky

Rene Magritte, The False Mirror, 1929.

The 1982 hit song Eye in the Sky by the British rock band The Alan Parsons Project did not – as it had often been claimed – allude to an Orwellian type of surveillance dystopia. Music geeks have settled on the likelihood that Parsons’ co-writer Eric Woolfson, a keen gambler who spent a lot of time in casinos, had been fascinated with hidden cameras watching his moves. These cameras, still referred to as ‘Eyes in the Sky’ in the industry, monitor suspicious gambling behaviour.

Four decades later, however, the song’s lyrics accurately capture a cold reality well beyond what its writers could have ever considered. Rather unsettlingly they reflect the impact of online digital surveillance processes driven by Artificial Intelligence (AI) that make their own rules and read our minds:

I am the eye in the sky

Looking at you, I can read your mind,

I am the maker of rules, dealing with fools,

I can cheat you blind,

And I don’t need to see anymore to know that I can read your mind. 1

AI-led surveillance is not just about facial recognition and the tracking of our movements in airports or during street protests. Nor is it just about self-drive cars, medicinal breakthroughs or robots taking over.

The digital realm – where we all interact with AI – allows us access to vast amounts of information and media platforms for the dissemination of our personal news and views. But it does so by systematically mining and exploiting what we as users unthinkingly provide – our raw data.

AI uses the information we gleefully submit about ourselves to classify, engage, cajole, persuade, change our behaviour and amplify our biases. It cheats us blind.

Behind casino cameras there are security people staring at monitors. No human eyes monitor the deluge of Facebook likes and shares, Google searches or tweets. Instead, there are predictive analytic processes, themselves devised by myriads of data scientists. These extract and suck in volumes of direct or indirect evidence of relationships of interest and apply so-called learning algorithms to figure out how to link, engage and then manipulate users.

Crucially, just as the cameras belong to the casino owners, these social media algorithms and processes also belong to someone – Big Tech platforms. The Mark Zuckerbergs and Jeff Bezoses. Those whom Harvard professor Shoshana Zuboff describes as Surveillance Capitalists2 : “[They] know everything about us, whereas their operations are designed to be unknowable to us. They accumulate vast domains of new knowledge from us, but not for us.”

To comprehend the magnitude of the threat it is useful to re-examine the term AI. The word ‘Artificial’ is deliberately camouflaged in a certain techno abstraction rendering it unbelonging and irresponsible. We say ‘Artificial’ as if it is independent, otherworldly.

If, instead, one uses the term Machine Intelligence (MI) one can begin to grasp that the machines physically exist, they are actually located somewhere and that they belong to someone – the Zuckerbergs and Bezoses.

It is important to identify this aspect because it is only then that owners can begin to be held accountable. Besides professor Zuboff’s warnings, there are what USC professor Kate Crawford says are consequences that relate to the natural resources involved to maintain these huge machine and data centres. There’s the incalculable consumption of energy and the pollution they produce; the labour processes that quietly exploit crowds of workers, human – not artificial. All these are on the ground not in ‘Clouds’ in the sky.

Zuckerberg and Bezos don’t care which film you liked, which book you bought, nor which football club or political party you support. But their clients do. Facebook has 3 billion active monthly users and 7 million active advertisers. It sells the data of the first group to the second fairly cheap but many times over. Our constant supply of information is resellable, reconfigurable, malleable gold.

All it wants is for you to keep sharing what you ate and where, what you read, what you bought, what you thought, who you voted for. It “maps out the graph of everything in the world and how it relates to each other”2 and sells that information to companies that then come at you thick and fast. Yet Mark or Jeff won’t allow you to look into their Machines and become quite tetchy when someone from within talks about what is happening inside.

LSE professor Damian Tambini speaks of the loss of human autonomy as a result of this increased capacity of smarter media to control the information available to us, by gathering ever more “granular” data about us. This, he says, has not only changed advertising, it is transforming politics. Democracy, he claims, faces a new vulnerability3. The entire information environment is not just being ‘managed’ for us, it is also being polluted with misinformation. Big Tech are doing the very minimum because they don’t want their business model disrupted.

Unless regulated, things will become darker and messier. The EU which is much better placed than the US and China to act has an obligation to do so quickly and decisively. Ethics, says professor Crawford, are necessary but not sufficient. What you see time and again, she says, is these systems empowering already powerful institutions – corporations, militaries and police4.

Clearly MI is a one-way street and can if regulated properly enable sustainable growth. But it is also important for citizens to become sensitive to – and alarmed by – how its deployment and misuse is beginning to infringe on their fundamental rights and the rule of law.

If you decide to listen to Eye in the Sky on YouTube check which song Machine Intelligence will recommend for you to click on next. Mine popped up the 1985 new wave track by Tears for Fears Everybody wants to rule the World.

1. Eye in the Sky –

2. The Age of Surveillance Capitalism – Shoshana Zuboff, 2019.

3. The Council of Europe Ministerial Conference Artificial Intelligence – Intelligent Politics held in Cyprus on 10-11 June 2021 explored these issues in depth:

4. Atlas of AI – Kate Crawford, 2021.