“We have entered the Intelligence Age,” proclaimed Sam Altman, the chief executive of OpenAI, in September. “Deep learning worked,” he explained, and this breakthrough in learning from data will unleash a wise era through which the more data becomes available, “the higher it gets at helping people solve hard problems”.
Altman joins other thought leaders, corporations like Google and Amazon, and organisations akin to the United Nations and the World Economic Forum, in pinning humanity’s hopes on higher information.
The logic is enticing. By harvesting all of the world’s knowledge, AI models can locate patterns, make correlations, and offer data-driven “insights”. The optimal solutions to our biggest problems are needles in an information haystack, so finding them exceeds the limited human mind. It is as much as technology like deep learning to “capture all of it”, analyse or train on it – after which offer up the sensible game-changing idea or most rational response.
Climate change, based on a Google report, could be simulated and alleviated using forecasting and modelling. Global conflicts, suggest AI engineers Tshilidzi Marwala and Monica Lagazio, could be modelled and mitigated.
But the recent US election showed the bounds of this rational framing of reality. Viral rumours and conspiracy theories (JD Vance and the couch, or “they’re eating the pets”) were gleefully shared. It seems some voters were motivated less by abstract policy and more by visceral disgust at those deemed different.
Humans usually are not perfectly rational and ethical. They are deeply emotional, factional and frictional – driven by feelings and friendships, fear and anger.
Donald Trump’s win was aided by tapping deeply into this darker and more “irrational” core of human nature, defying the polls. It was never about perfect information.
In the past five years, my research has explored how technologies construct knowledge – but in addition exploit emotion and amplify radicalisation. To understand the present political moment, we’d like to know each the bounds of reason and the facility of unreason.
Climate change, genocide and data
Intelligence is a dead end. The entangled social, political and environmental crises we now face is not going to be addressed by having more information.
Climate change is Exhibit A. Since 1990, the Intergovernmental Panel on Climate Change has published six assessments detailing the drivers and impacts of world warming. The latest report is 3,949 pages and was based on greater than 14,000 scientific papers.
Despite this deluge of knowledge and expert evidence, the planet has already passed the 1.5°C threshold above pre-industrial levels and will pass 2°C by the tip of the last decade. Indeed, over the three many years of the United Nations’ Conference of Parties meetings (1994–2024), carbon emissions have steadily ticked upwards. A mountain of facts has not budged the voracious extraction and consumption of “business as usual”.
Gaza is Exhibit B. Law for Palestine, a not-for-profit, youth-led legal organisation, has meticulously compiled a database with greater than 500 instances of incitements to genocide.
Last week, South Africa filed its Memorial to the International Court of Justice, a key document in its case against Israel. Its evidence is detailed in greater than 750 pages of text, supported by exhibits and annexes of greater than 4,000 pages.
And yet Israel’s assault on Gaza continues, with deaths mounting up. According to Al-Jazeera, among the government’s kill lists are generated via big data evaluation, a type of AI-assisted genocide.
In each of those examples, intelligence was ignored or just co-opted to rationalise a desired motion. As a George W. Bush adviser famously stated: there are individuals who imagine solutions emerge from a “judicious study of discernible reality” and there are those that “create our own reality”. While analysts are analysing, empires are acting, reshaping the world.
Facts could be trumped by alternative facts. Scientific reason could be ignored or refused. Machine learning could be overpowered (or weaponised) by men with guns.
In Trump’s first presidential term, I analysed how Immigration and Customs Enforcement had begun to make use of Palantir technologies, which assembles data into a strong visual interface to help in deportations.
Even a decade ago, the agency was scanning three billion pieces of knowledge, from licence plates to border-crossing data, using ostensibly sophisticated software to discover targets and residential in on “illegals”. And yet these data-driven insights were ultimately a pretence: tokenistic evidence for a war on immigrants the Trump administration had already decided to wage.
Any claims that this computational logic was “post-racial” quickly disappeared within the violence of deportation raids, when immigration officers punched individuals within the face, ordered them to the bottom, placed boots on their head, and referred to them as “Mexican shit”.
Machine intelligence, ethics and morals
“Enlightenment reverts to mythology,” predicted influential cultural theorists Theodor Adorno and Max Horkheimer 80 years ago in Dialectic of Enlightenment.
A society obsessive about the perfection of reason inevitably breeds its opposite: brute force and proud irrationality. Their insights were removed from abstract. The German theorists had fled for his or her lives, penning their groundbreaking classic in exile in California while Hitler’s murderous regime played out across the Atlantic.
Reason, the duo observed, has morphed from its ancient origins to change into an approach and a set of techniques. Reason is about ordering, classifying, and deducing from data – and applying this “abstract functioning of the pondering mechanism” in an appropriate or scientific way.
Such techniques sit on the core of latest machine learning systems, which train, fine-tune and generate claims from data.
This makes reason free-floating, a set of operations that could be applied to anything and the whole lot. Such “intelligence” is effectively content-agnostic: devoid of any intrinsic aim or purpose. Reason doesn’t (and indeed cannot) judge the moral or moral implications of its use.
But if “all affects are of equal value”, the duo note, then survival becomes “essentially the most probable source of maxims for human conduct”. Self-interest is essentially the most “reasonable” approach of all, and anything that hinders this — social cohesion, protections for the marginal or vulnerable, contributions towards the general public good —should be disbursed with.
In a zero-sum game, giving anything to “them” simply means less for “us”. Within nations, we see this in polarisation and division. Internationally, we see it in hardened borders and bunker mentalities. Any other strategy besides domination is suboptimal.
The ticking time bomb of reason
Reason creates a type of ticking time bomb on the core of society. Reason is a versatile amalgam of knowledge harvesting, data-driven decisions and optimised operations that wins votes and attracts investment.
But because reason is uncoupled from ethics, it may possibly and ought to be applied to anything: no aim is best or worse than some other. The result for Horkheimer and Adorno is a type of moral relativism.
An ideal example is Robodebt, Australia’s use of knowledge analytics and machine learning to halt welfare payments. This mechanism for “automating inequality” created untold misery for individuals who were already vulnerable. Denying care became procedural and due to this fact “reasonable”.
This calculative logic is each practical and unassailable. Once rooted within the national consciousness, it becomes tough to dislodge. Actions not tied to survival are hard to defend; values circuitously linked to self-interest are difficult to rationalise. Any connection or responsibility to the thousands and thousands of others within the country risks breaking altogether.
In the tip, this national community can only be held together by sheer force or terror, Horkheimer observes: liberalism tilts over into fascism.
The rise and rise of strongman leaders, poised between fascism and popular nationalism, exemplifies this tilt.
In Hungary, authoritarian leader Viktor Orbán has a 14-year grip on power, undermining judicial independence and press freedom together with his brand of self-styled “electoral autocracy” or “illiberal democracy”. In Russia, Vladimir Putin has created a cult of power, annexing Crimea, waging war in Ukraine — and crushing protest with violence and imprisonment.
Most recently, Trump has over again risen to power, vowing to punish the “enemy inside” and perform mass deportations. “I’m your warrior, I’m your justice,” he vowed, “and for individuals who have been wronged and betrayed, I’m your retribution.”
The arc of reason eventually arrives at a brutish world ruled by essentially the most brutal. Democratic “civilisation” collapses back into domineering barbarism.
Age of intelligence or age of anger?
As reason’s grand guarantees inevitably collapse, people grow disenchanted or disaffected, latching onto regressive worldviews that make the world make sense.
My own work has explored how online platforms have repackaged racist, sexist and xenophobic ideologies into alluring recent types of hate, spawning incels and edge-lords, Christian nationalists and QAnon conspiracy theorists.
While such prejudices can actually be condemned, they make clear our world today. For those that adopt them, these narratives offer a compelling account of how the world works, why someone is stuck – and who or what’s responsible.
New studies have shown that white Americans who perceive themselves to be in “last place” within the racial status hierarchy are most drawn to alt-right extremism. The undeniable fact that this perception is divorced from reality — that when it comes to education, income and imprisonment rates they continue to be at or near the highest — doesn’t diminish its power.
The power behind these narratives just isn’t logical, but emotional. This just isn’t the liberal subject, rigorously weighing the facts before selecting the claim that best conforms to empirical evidence and contributes to the general public sphere. No, it’s about grievance, loss and a way of betrayal by the powers that be.
This affective power is echoed by sociologist Arlie Hochschild, who spent five years within the Republican stronghold of Louisiana, where many suffer from poor health and lifelong poverty, within the years preceding Trump’s first election. After countless conversations with residents, she stresses that their story is one among anger and mourning: a “feels-as-if” story that bypasses judgement and fact.
For the white subject, the shock and sense of loss is tangible. Advantage just isn’t just “slipping away” but is being “taken by undeserving others”. Having stood atop the social hierarchy for thus long, the lack of stature seems like a tumble into the abyss. Regressive views akin to hyperconservatism or ethnonationalism offer a handhold or lifeline. And yet, if they carry a renewed sense of stability, in addition they cultivate a deep-seated hostility.
In this sense, our moment just isn’t the Age of Intelligence however the Age of Anger. This just isn’t to condone naked violence, baseless hatred and shameless propaganda, but to recognise the bounds of reason in contemporary life.
The irrational human animal
“If there’s anything unique in regards to the human animal,” philosopher John Gray observes, “it’s that it has the power to grow knowledge at an accelerating rate, while being chronically incapable of learning from experience”.
For modern humanists, humans are inherently rational. The primary future challenge is merely tips on how to make them more logical, more civil, more reasonable. This is the human that Altman and other techno-positivists keep in mind after they suggest AI will solve our problems. It is barely a matter of augmenting human intelligence with machine intelligence.
But history suggests our crises and responses have all the time been shot through with the “irrational”, a darker mélange of emotional power, bodily violence and political will.
Intelligence — even scaled, automated and operationalised by artificial intelligence — is not going to save us.