Big data, big trouble

The development and increasing use of artificial intelligence and big data gives rise to many more complicated issues than just the impact on the jobs market.

The invention of steam power and mechanisation of manufacturing and farming in the 18th and 19th centuries helped to enable the abolition of slavery, then child labour and then the introduction of the eight-hour work day.

The Industrial Revolution didn’t herald a new era of leisure so much as respect for human dignity and freedom, a proper childhood and education, and evenings at home.

It was all good, on the whole. So what will the current revolution enable? What horrors can we abolish this time? Or will we just get afternoons at home as well?

Much of the analysis of the future of work these days bears a passing resemblance to the views of Ned Ludd and his followers who were smashing machines to protect jobs in England the early century.

Happily no one’s trying that again, but merely focusing on employment is far too simplistic: there is growing disquiet that in other ways technology is moving more quickly than humanity’s ability to deal with it.

An open letter has just been published by the Future of Life Institute signed by nearly 1500 influential thinkers, academics and entrepreneurs – including Stephen Hawking, Erik Brynjolfsson, co-author of Race Against the Machines, and Elon Musk, CEO of Tesla – calling for more research into the impact of artificial intelligence.

“Because of the great potential of AI, it is important to research how to reap its benefits while avoiding potential pitfalls.

“…our AI systems must do what we want them to do.” 

A longer paper published simultaneously goes more deeply into some of the issues, such as machine ethics (who is responsible when a driverless car crashes?), whether economic statistics need to be rethought, and how to deal with under- and unemployment.

“Unemployment is not the same as leisure, and there are deep links between unemployment and unhappiness, self-doubt, and isolation…”

Of course the issues of employment are already upon us, well before driverless cars take over the roads, and right now the past and future are coexisting.

I was at Bunnings over the weekend and checked out at one of the self-serve checkouts: no human required. It’s the same at the supermarket. Then I went over the road and bought a print for the wall: one of a small team of shop assistants laboriously wrote the details of the purchase into a book, longhand, before processing the sale.

It’s pretty clear that shop assistants will go the way of service station driveway attendants; self-serve checkouts work fine, and they won’t just be used in supermarkets. A lot of people are shop assistants, but stores that survive online retailing won’t be employing them, or at least not many.

Over the sort of time frame that covered the Industrial Revolution – roughly 100 years – a lot more jobs will disappear, but more significant issues than employment are starting to come up.

For example, the Financial Times this morning published research from the Universities of Cambridge and Stanford into how well Facebook now understands its users.

The answer is that by using “likes” on Facebook, a computer can understand a person better than his or her siblings and spouse.

In fact 'big data' records everything we do. As this is put together and analysed by AI software, we will not only be unemployed we will be entirely laid bare for governments and companies to watch us, and toy with us.

While machines, including robots, algorithms and software, can do most things more efficiently than humans and don’t need to rest, there is much more going on than simply replacing repetitive work with machines and bots.

It means parallels with what was happening 200 years ago are limited and misleading.

Once the Luddites were locked up and dispersed around 1820, it was full steam ahead, so to speak, with improving productivity through innovation.

Moral and ethical issues were few, and perhaps the only non-industrial challenge was deflation. Prices fell consistently through the second half of the 19th century, accompanied by frequent financial crises – leading eventually to the creation of the Federal Reserve and central banking in 1907.

This time around we have deflation again, although this time there is a phalanx of central banks combating it, albeit with limited success (last night the yield on Japanese five-year government bonds fell to zero).

There is also the same sort of employment challenge this time as there was 200 years ago.

So what will be the modern equivalent of abolishing slavery and child labour, not to mention the introduction of the eight-hour working day? I suspect that boring repetitive work itself will be abolished. In 200 years, people of the future will read in wonder about how humans sat at a production line doing the same thing over and over, or keyed data into a computer.

But this time there are also much more complicated issues at stake as well, thanks to artificial intelligence and big data.

Who controls my data? Who has access to it? What are we going to let machines do?

Just read this rather convoluted section from the Future of Life Institute’s report:

“If an AI system is selecting the actions that best allow it to complete a given task, then avoiding conditions that prevent the system from continuing to pursue the task is a natural subgoal (and conversely, seeking unconstrained situations is sometimes a useful heuristic).

“This could become problematic, however, if we wish to repurpose the system, to deactivate it, or to significantly alter its decision-making process; such a system would rationally avoid these changes.”

I recently watched (again) a film called “Eagle Eye” in which a ‘system’ tries to kill the US government because it heuristically saw a threat to deactivate it.

Great movie, but impossible? Apparently not.

Related Articles