The age of specialised AI

Could a computer do your job?

That’s the first time I’ll ask you that today. The second will be in about three minutes’ time. Let’s see if your answers vary.

This is something I’ve been thinking about and researching for a few months now. I’ve spoken to some interesting people about it (more on that another time). But the question occurred to me particularly on Saturday.

I was sitting on the plane ready to fly out here (I’m in California this week), looking idly out the window as the last passengers boarded. Outside, two blokes were stood in the rain loading the cargo onto the plane.

And when I say ‘stood there’, I’m not suggesting they were being lazy.

All the heavy lifting was being done by a kind of oversized robotic arm that picked up each crate, twisted it around in mid-air and slid it into the hold. It looked like a gigantic game of plane cargo Tetris. The two men were overseeing things. One was operating the machine. The other was watching him (presumably for safety). Whatever they were talking about was obviously incredibly interesting, because it looked like they barely needed to be there.

It made me think back to when I used to go on holiday with my parents, and the same job would be done by a team of people, loading the bags by hand in a kind of packing frenzy. There’d be swarms of them.

These days it’s two people. And to be honest, they didn’t look like they were exactly irreplaceable. The real heavy lifting (excuse the pun) is all done automatically by machines.

OK, this isn’t exactly a new phenomenon. Automation and machinery has been displacing human labour from the workforce for centuries. That’s where the term ‘Luddite’ comes from. The story goes that in 1779 a guy called Ned Ludd smashed two knitting frames in a textile factory as a protest against machines replacing people in manufacturing.

It was a futile gesture.

But perhaps it’s one that will become more and more relevant in the next decade or so.

You see, one way of thinking about the Industrial Revolution is that it meant humanity was no longer constrained by the strength or speed of any particular man or woman. Mechanisation allowed us to harness a power many times greater than our own individual strength. It surpassed the physical limitations of our own muscles.

Over time, that’s led to machines replacing humans in an increasing number of professions that rely on physical labour.

The digital revolution, on the other hand, has allowed us to break the constraints not of our muscles, but of our brains.

Computers can increasingly out-think, out-compute and out-perform people in a growing number of white collar ‘intellectual’ (ie based primarily on non-physical skills) professions.

An example. Having watched the cargo being loaded onto the plane, I opened the newspaper and came across a story called “Dogs Might Fly: Ground-breaking TV experiment will train a Labrador to become a pilot”.

From The Independent:

“According to an aviation joke, planes only need two crew members – a pilot and a dog. The pilot’s job is to feed the dog. The dog’s job is to bite the pilot if he touches the controls. But now a ground-breaking television series will place Britain’s most intelligent canines in the cockpit in a bid to discover if a dog can be successfully taught the skills to fly a plane.

“Using participants handpicked from rescue centres, the Sky 1 series, Dogs Might Fly, aims to prove that the memory and reasoning abilities possessed by the brightest pets could be directed towards mastering the controls of a light aircraft.”

I’ll admit, if I was boarding a plane and saw an air hostess take a bowl of chum into the cockpit, I’d probably think twice about flying.

But the piece made me think: we’re perfectly happy to accept that the pilot’s role on a flight is to oversee the autopilot. Essentially, the plane is flown by a computer. A non-human. A robot.

In fact, autopilot is essentially a form of ‘narrow’ artificial intelligence (AI). By narrow, I mean it is brilliant at a given task, but useless at everything else. For a reference, the other kind of AI is ‘strong’ AI – something with human intelligence and the ability to turn its (metaphorical) hand to a huge variety of tasks.

We’ll talk about this in more detail another time. (Right now there’s a huge amount of debate about when a strong AI will be created and what will happen when it is.)

But in the context of what we’ve been talking about, ‘narrow’ AI is a much more pressing issue. Narrow AI (I’ll call this NAI from now on) is here. And it’s going to change the world radically.

The key to NAI is specialisation. We can now create computers that can out-think and out-perform humans in a huge and growing number of tasks.

No, you can’t have a conversation with one, as you’d be able to with a strong AI.

But NAIs are already replacing humans in all sorts of places. They can outdrive humans (as the ‘driverless car’ is proving – there have been something like 100,000 hours’ worth of tests without an accident). They can recognise language and organise your day (SIRI on an iPhone), translate one language to another, make recommendations for a book you might enjoyThey can even write articles, research legal case precedent and diagnose illnesses. From The Time’s Raconteur supplement earlier this year:

“AI-based systems are already making inroads into knowledge-based industries such as journalism. The Associated Press news agency plans to automate the writing of corporate earnings reports with an AI system called Wordsmith, which spots patterns and trends in raw data and then describes those findings in natural language. Similar systems could produce legal documents, carrying out many of the tasks given to paralegals and junior associates. Due diligence and litigation discovery software is already highly developed.”

The kind of strong AI we see in films might be a long way away. And narrow AIs don’t fit the image most people have of super-intelligent machines (most people seem to imagine a compassionate robot like C3PO in Star Wars). But huge numbers of highly specialised and adept NAIs are disrupting the way white-collar professions operate. How long will it be before they start displacing the workers themselves?

What’s the modern day equivalent of Ned Ludd? Try as they might, a journalist can’t smash a writing algorithm up to make a point.

Well, they could have a go. I’d actually quite like to see what that looks like.

But it’d be just as futile as Ludd’s protest 230 years ago.

As John Stepek pointed out in last week’s podcast, this actually goes all the way back to Adam Smith’s Wealth of Nations. One of Smith’s key ideas was the role specialisation plays in the division of labour:

“The greatest improvement in the productive powers of labour, and the greater part of the skill, dexterity, and judgment with which it is anywhere directed, or applied, seem to have been the effects of the division of labour.

“To take an example, therefore, from a very trifling manufacture; but one in which the division of labour has been very often taken notice of, the trade of the pin-maker.

“A workman not educated to this business (which the division of labour has rendered a distinct trade), nor acquainted with the use of the machinery employed in it, could scarce, perhaps, with his utmost industry, make one pin in a day, and certainly could not make twenty.

“But in the way in which this business is now carried on, not only the whole work is a peculiar trade, but it is divided into a number of branches, of which the greater part are likewise peculiar trades.

“One man draws out the wire, another straights it, a third cuts it, a fourth points it, a fifth grinds it at the top for receiving the head; to make the head requires two or three distinct operations; to put it on, is a peculiar business, to whiten the pins is another; it is even a trade by itself to put them into the paper; and the important business of making a pin is, in this manner, divided into about eighteen distinct operations, which, in some manufactories, are all performed by distinct hands.

“I have seen a small manufactory of this kind where ten men only were employed, and where some of them consequently performed two or three distinct operations. But though they were very poor, and therefore but indifferently accommodated with the necessary machinery, they could, when they exerted themselves, make among them about twelve pounds of pins in a day. There are in a pound upwards of four thousand pins of a middling size.

“Those ten persons, therefore, could make among them upwards of forty-eight thousand pins in a day. Each person, therefore, making a tenth part of forty-eight thousand pins, might be considered as making four thousand eight hundred pins in a day.

“But if they had all wrought separately and independently, and without any of them having been educated to this peculiar business, they certainly could not each of them have made twenty, perhaps not one pin in a day; that is, certainly, not the two hundred and fortieth, perhaps not the four thousand eight hundredth part of what they are at present capable of performing, in consequence of a proper division and combination of their different operations.”

In short, breaking one job down into a series of highly specialised positions is the fastest way to increase output.

The same is true of NAIs. A computer may not yet be able to replace a journalist. But break the job down into its constituent parts and you see that many of the smaller parts of the job could easily be done – and potentially done well – by an NAI.

Research

An NAI could easily read and analyse more information on a given subject in a given period of time than a human. A specialised NAI could crunch more data and create more charts (for instance) than a human. A different NAI could comb thousands of articles for quotes and information. These processes are more or less the same for fact checking.

Writing

This is close to the bone for me. I’d like to think a computer couldn’t “out-write” a human in the creative sense. But in terms of grammar, sentence composition and the like, computers are already our equal. And as John pointed out last week, an NAI could easily be programmed to scan for regularly over-used and clichéd phrases and automatically avoid them.

Editing

Again, a specialised NAI – with no bias or personal opinion one way or another – could easily cut down a piece of writing to a given size. They’d do it quicker. Things like mistakes, grammatical and spelling errors, etc, would all be cut out more efficiently. And there’d be less arguing. Hard for a disgruntled writer to bully a sub-editor with no emotions.

You get the picture. The overall role may be far too complex for one NAI to replace. But break a role down into a series of specialised positions and you start to see the future.

Don’t get me wrong. I’m not portraying this as some sort of dystopian nightmare in which entire professions disappear. The rise of specialised NAIs will almost certainly bring about incredible opportunities (particularly for the people who invest with all this in mind: more on that other time). But no one likes radical change. And any profession under threat will always push back. Just ask a London cabbie what he thinks of Uber.

In the meantime, I’d like to know what you think of all this. You may think it’s a load of old tosh. Or perhaps not. You can reach me at [email protected].

Before you write to me though, read this piece over one more time, think honestly and ask yourself: could a computer do your job?

 

Category: Investing in Technology