Quantosaurus Rex: Rewards & Risks

Humans are vanquishing uncertainty in all areas, but we don’t know what comes next, says The Age of Prediction: Algorithms, AI, and the Shifting Shadows of Risk by Igor Tulchinsky and Christopher E Mason . Our world is wired by billions of predictive algorithms, more powerful, ubiquitous and precise than ever. We model markets, disease and traffic; countless fields are being transformed by these tools.

Oceanic amounts of data, more computational power and sophisticated analytical processes with machine learning and AI, are making enormous things possible. Two forces shape our life: the light of prediction and the shadow of risk, says the book.

Prediction in stationary systems can approach perfection; large numbers drive certainty to near 100%. Datageneration has taken off like a rocket in recent decades. Digital data is currently estimated at 10 zettabytes (100 followed by 21 zeroes) – too much to wrap our minds around – and it’s growing exponentially. Smart thermometers can help countries detect disease spread, shipping data tells you about migration movements, purchasingdecisions can tell you aboutabrewing war, and phone location data,of course, can be put to all kinds of uses.

Complex non-linear systems had been harder to crack statistically: the way a mob behaves or consumer choices, or what a turbulent stream of water does, or the mutations that drive evolution, are seemingly random. But increasingly, even this complexity can be penetrated. Financial markets were revolutionised by modern portfolio theory that aims to quantify risk and reward, and AI and machine learning have changed that game. The explosion of vaccine work that emerged after Covid relied heavily on machine learning.

The book is clear-eyed about the enormous power of these technologies, and their current limits. Artificial intelligence is an imprecise term that includes machine learning, deep learning, natural-language processing, expert systems and fuzzy logic. The history of AI is a ‘graveyard of famed predictions’, but at root, it is simply the idea that evolution could expand beyond biology.

More powerful predictive capacities do not mean better outcomes, though. In warfare, AI and robotics threaten to escalate risk. What’s more, behavioural responses to changing risk have always been fluid. When individual risk is reduced, behaviour grows more reckless, as is seen in insurance, for instance. Predictive technologies spawn unintended consequences – it can increase longevity through healthcare, but it doesn’t tell you what to do with older populations. Risk moves around like a game of whack-amole, popping up in new forms even as older ones are subdued.

Perfect accuracy is not here yet, but what happens when we come close to it? Absolute prediction is repressive; the future shrinks to one inevitable option. There are other disturbing possibilities – imagine genetically engineering people for the right job. Pre-election polls can corrupt democracy if voters believe they are infallible. Big tech companies with their minute surveillance can commandeer our actions.

Meanwhile, look at how we face the existential threat of climate change – its impact can be modelled, risks can be quantified, behaviour can be modified, tech can be mobilised. But the human and social snarl of finding an effective response carries on. Ultimately, for all the power of our tools, the book calls for a certain humility, an awareness that there are things beyond our imagination, let alone prediction.

END OF ARTICLE