Brave new world: rise of the algorithms

Technology: Hello World, Hannah Fry Doubleday, hardback, 256 pages, €19.60

Plugging in: Fry suggests we form a complementary relationship with algorithms

Plugging in: Fry suggests we form a complementary relationship with algorithms

Hello World

Plugging in: Fry suggests we form a complementary relationship with algorithms

September 09 2018 02:30 AM

Since probably the dawn of civilisation, people have had an ambiguous relationship with technology. As a collective, we generally embrace it and appreciate its value, how it makes our lives longer and safer and better. As individuals, though, we're often uneasy about technological change - especially the speed at which it happens - even if we're not always entirely conscious of this unease.

This seeming contradiction lies at the heart of Hello World. Subtitled 'How to be Human in the Age of the Machine', Hannah Fry's book explores algorithms and how they have radically altered human society - sometimes for the good, sometimes not so good.

As well as being a popular TED talker, broadcaster, podcaster, public speaker and author, Fry's day job is as a professor in "the mathematics of cities" at University College London, where she "uses mathematical models to study patterns in human behaviour".

This solid grounding - in facts, statistics, provable and measurable phenomena - ensures that she takes an admirably balanced approach to her subject. Fry is not ideological about the digital world, which makes for a much more enlightening book. Hello World is neither one of those 'machines will make you happy' pieces of tech-guru propaganda, nor is it an overly pessimistic diatribe about how the robots are just waiting for their chance to enslave humanity. (For one thing, as Fry points out, the dawn of properly self-aware AI - if it ever even happens - is a long way off.)

Algorithms undeniably improve things in a lot of ways. But there are, naturally, attendant problems. Hello World examines both sides, section by section: political power, data, crime and punishment, medicine, travel and art/culture.

The most 'important' of these, I would imagine, are medicine, politics and crime; the most interesting for me, on a personal level, was artistic creation. First, though, Fry explains what an algorithm is.

It's a little tricky to define, she admits, and not everyone agrees on the parameters. But essentially, an algorithm is a set of instructions. Technically, of course, this means that the bit of paper which came with your Ikea Billy bookcase can be considered an algorithm; in practice, the term refers to computer software.

There are, very broadly speaking, two main types: one is more or less pre-programmed and left at that; the other 'learns' as it goes along. (That's the one which will eventually mutate into a conscious intelligence that launches World War III - or not, as the case may be.)

And these clever little bits of coding have been used by people in a staggeringly wide variety of ways. Every time Netflix recommends a movie or TV show, that's an algorithm. Supercomputer Deep Blue defeating Garry Kasparov in a famous chess game in 1997: that was an algorithm. Cruise control and sat-nav on your car: algorithms.

This is relatively trivial stuff. More serious -and this, Fry repeatedly insists, is something society and politicians will eventually have to grapple with - are those big issues of justice, medicine, elections.

In a world where prison sentencing and parole are decided by computers, or where life insurance is denied because an algorithm identifies that someone has a certain probability of contracting Huntington's disease, these are life-and-death matters.

She poses a series of questions, the digital equivalent of those 'moral quandaries' we enjoyed mulling over in school civics classes. How much can social media influence a vote - and should we restrict the freedom of its users and providers? Would you agree to your entire medical history being publicly available if it resulted in major breakthroughs and cures? In the event of an impending collision, should self-driving cars be programmed to save the life of the passengers or the pedestrians?

Ultimately, I suppose, it all boils down to one question: how much should we trust algorithms? Fry points out that perfection is impossible, for a machine as much as any of us, so there's no point in lambasting an algorithm that doesn't work just right 100pc of the time.

She suggests a sensible solution: forming a complementary relationship with them. Use the processing power and attention to detail of algorithms, paired with the intuitive smarts and abstract thinking of people.

Video of the Day

Most importantly, keep the human element at every step of the way. Eschewing the use of algorithms entirely is something akin to cutting off your nose to spite your face; trusting them fully and blindly is reckless and stupid.

Hello World is a consistently interesting, nicely written and accessible primer for anyone who needs to get to grips with this brave new world of algorithms - which, in the end, is all of us, whether we want to or not.