Posted by laura on

It’s not about human vs machine

I find myself talking a lot about automation, both as part of the talks I give and in answering the questions that follow. It’s a polarising issue. Some people believe that the robots will replace us all. Others are dismissive of the idea. Each side cherry-picks facts and ideas that support their belief, or desired headline.

As usual the reality is somewhere in the middle. But that doesn’t mean it will be any less devastating for many people — or any less advantageous for others. Because the battle is not between human and machine. It is between humans augmented by machines, and those that aren’t.

It’s hard to make this argument without sounding like an old-school class warrior, so note that I imply no judgement in the following statement. But the reality is that this is about capitalism.

In a capitalist system organisations and their leaders are incentivised towards decisions that maximise profits. In our current iteration of capitalism, we incentivise profit creation over a relatively short term, performance being measured year-on-year or quarter-on-quarter. Any opportunity to increase that performance is generally taken, sooner or later. The companies that don’t take those opportunities are displaced by competitors that do, at an accelerating rate.

Many tasks can be accomplished by machines at a lower cost per unit. They may lack many valuable human characteristics but when measured against short term profit metrics, the machines often beat humans by many orders of magnitude.

The exponential falls in the cost of technology, and particularly information technology innovation inside organisations, means that more companies can start to experiment with more automated ways of doing things. When those experiments work, they have a competitive advantage. Which enforces more rapid adoption on the rest of the market.

Put simply, wherever machines offer an increase in profitability, they will be deployed. The caveat to this is those organisations with a social motive or a more long-term perspective on value creation, who can afford to challenge the status quo. But these organisations are not the majority.

The result is that many human jobs are naturally taken by machines. Or to put it more accurately, one human being augmented by machines can do the work of five, ten or more others.

Take the call centre as an example, a sector that employs around a million people in the UK according to Unison. Imagine that each one of the 5,000 call centres in the UK introduces an artificial intelligence agent on its front line. Not the ‘Press One to lose your mind’ systems we have had to date, but conversational, smart systems that can interact in a near-human fashion. Imagine no queues: every call answered in two rings. Imagine 80% of queries answered first time — an estimate from existing AI customer service deployments in the US. Even once you add a generous number of staff to maintain the AI, that’s probably two thirds of the jobs gone. You’ll still need second line support for the queries the machine can’t answer. But because it learns, that proportion falls every day.

To a greater or lesser extent, the reality is the same across many industries. It’s not all the human jobs that get replaced. Rather a smaller number of humans can do the work of many. It will be the same in every industry from logistics to law, accounting to advertising, manufacturing to financial markets. New jobs will undoubtedly be created, but it’s hard to see them being in the same volumes as those that are automated.

Is this sustainable? No, I don’t think so. At least not in our current economic system. But it will take time for that system to change and adapt — much longer than it takes for companies to adopt these technologies. So while the incentives remain the same, we will see mass automation. Or perhaps more accurately, mass augmentation.

And that presents a whole new — or arguably very old — set of problems.

Share this article ...

Tom Cheesewright