Main content

Skynet is here. We just don't know it.

Den Howlett Profile picture for user gonzodaddy November 12, 2017
AI is already with us even if we don't know it. The problem is the manner in which it impacts people's lives through the relentless drive to efficiency. Intended or not, change is urgently needed.

In the polemic Terminator movies, the most repugnant form of AI takes over the world, replacing humans with machines. While the science fiction of those movies lives long in the memory, the consequences and unintended consequences of AI are already being felt. In that sense, Skynet is already here. Anyone using Google, Facebook, Twitter, Airbnb, Ocado, Apple's Siri and/or Amazon (the list goes on) is already using some form of artificial intelligence. In some cases, services are delivered with the aid of robots. raising the topic, we've been here before.

In May, Phil Fersht wrote a polemic about the post-AI world to which I responded. It must be that time again because Fersht is back at it. He is unashamedly upfront about the impact of AI on the technology workforce. In his latest gut check, Fersht says - and I apologize to him for chewing out a big chunk of his prose but it's important. (Emphasis added.)

Let's just make the call - AI is indirectly and inextricably tied to the elimination of "unnecessary" labor, by nurturing systems that get smarter with each incident and transaction. The smarter and more autonomous your operations become, the more agile and efficient your business becomes. That's not a terrible thing - in fact, you are doing your valued staff a huge favor by keeping them employed and keeping them relevant to your business.  But you are not creating a net influx of jobs into your organization, you are becoming more fluid and competitive.

Sure, you'll probably look to add some Python and R developers, Machine Learning experts, serious data geeks and design thinkers - or you may just pay consultants to do it all for you - but the bottom-line, here, is that you're going to be shedding a lot of your left-brained staff performing jobs that can be artificially automated, at a much faster rate than you'll be adding the data-oriented people you need to digitize your business.  AI is about doing more with less, not more with even more - let's get real.

Don't get me wrong, the possibilities of faster, smarter, touchless data flows between the customer and the operations of the business, are critical to promote competitiveness and survival, but let's stop sugar-coating the true purpose of data driven intelligence - the less businesses need to rely on people and the more autonomously they can run processes, the more nimble and profitable they will become.  Now if these businesses then choose to reinvest their new-found wealth hiring loads more people, I will tip my hat to these purveyors of job hope, but let's fact facts, the companies of the future will be running a lot of smart technology with a smaller group of savvy people to manage it all.

Business always wins

Anyone see the flaw in the argument? What are you going to do with all the people that have been displaced? The answer is self-evident. Unless they can be retrained to undertake right brain work, then they end up on the economic scrap heap. So far, so Taylorian. And still no clear answers.

But as has been said on many occasions, it's not the fact these displaced people go away, it's that there is no obvious correlation between a reduced workforce and increased profit. Why? While employed, these people are consumers. When unemployed, they become recipients of some kind of welfare. In turn, that means the amount of 'stuff' that businesses can sell shrinks. And if business shrinks then margins erode as the price of goods falls.

The net effect in a situation where automation really does improve efficiency is that any short-term gains are wiped out by market erosion of the goods being sold by virtue of the fact there are fewer people able to buy goods except at new, discounted prices.

Fordism to the rescue?

It's not rocket science. In my view, arguments fall apart when failing to anticipate what happens to demand. In that, I believe we should roll the clock back to the time when Henry Ford was innovating his automotive production line and dealership network.

Ford's great insight wasn't that you could do more with less but that if you could do more, pay the workforce more and keep the factory gate price at a point where demand is stimulated, then everyone's a winner. Ford wasn't building a production line to sell to a mythical American public. He was building cars as fast as he could to first sell to his own workers and then to the rest of middle America.

It was a profitable business to the point where according to Wikipedia, Ford's estate was worth $188 billion at 2008 values. To put that in context, Jeff Bezos, currently thought to be the wealthiest person in the world is valued at a mere $95 billion.

How much and when?

The real question is one of quantum. I've said it before, but I remember when robots were introduced into the automotive industry. It's been an ongoing story for 35 plus years. At first, the effects were small but at certain points, entire plants were closed. Fortunately, many of those displaced were either close to retirement or readily found other work. Will the same be true this time around? It is hard to know. What is of greater concern though is not the individual industry impact but the overall impact of advances applied to many industries at the same time.

In Fersht's analysis, the BPO market is likely OK for the next five years, but it is the five years after that when displacement becomes noticeable. That's a long time in technology terms and, I'd argue, is sufficiently long for us to see fresh breakthroughs that could change the course of predicted history. In other industries, we might well see much faster change. Check this from John Mauldin:

There is a new robotic machine called an Iron Roughneck that reduces the human labor required to connect pipe from a crew of 20 down to a crew of five. And those jobs were quite high-paying. Here’s a picture of this new robotic roughneck. Fifteen workers per site at well over $100,000 a year each? Does that machine look like it cost more than a few million? I bet it amortizes pretty quickly, and that’s why it is being rapidly adopted….. The amazing thing is that this transformation happened in two years; it didn’t take a generation or even half a generation.

The good side of AI

The flip side, of course, is that while Amazon is held up to be largely responsible for bricks and mortar retail demise, it has created many jobs in the process. This from the FT in July:

The jobs created are harder to pinpoint. Amazon employs fewer than three times as many people in the US as Walmart did in 1985, when its founder Sam Walton was crowned America’s richest person — despite having five times the inflation-adjusted revenue. But whereas Walmart ran most of its logistics in house, Amazon relies extensively on parcel carriers and agency workers; UPS alone has added 100,000 jobs in the past 16 years.  Officials do not count these as retail jobs. But Michael Mandel, an economist at the Progressive Policy Institute in Washington, reckons that once warehouse workers are taken into account, new ecommerce jobs outnumber brick-and-mortar losses by 54,000 over the past year.

Elsewhere? The FT continues:

Ocado employs 12,000 drivers, warehouse operatives and other staff to deliver perhaps £1.6bn worth of groceries, including the estimated revenues of a website the company runs for rival Wm Morrison. That is 76 staff for every £10m of retail sales, compared with just 36 to generate the same revenue at Aldi’s UK stores.

Ocado's not stopping there. Check this:

Ocado Group Plc put together an order of 50 items, including produce, meat and dairy, in five minutes. Fulfilling a similar order at one of the company’s older facilities takes an average of about two hours. The secret: a fleet of 1,000 robots that scurry about a warehouse snatching up products and delivering them to human packers.

Consumers who use these services love the convenience and efficiency, but the bigger question is whether the outcomes for those who work for Amazon, Ocado et al are meaningfully better than the alternatives.

In my earlier story, I argued that:

...there needs to be something of a reset in understanding how economies operate, the responsibilities of mega wealthy businesses and the families that control them and a setting aside of the ‘left-right/socialist-capitalist’ labeling that stifles otherwise useful debate.

To that extent, I would prefer to see business get behind Mark Zuckerberg’s invocation to ‘purpose,’ something to which Fersht alludes. It’s easy to snicker at someone who has made it and has the luxury of being able to say whatever he wants, but then that is to deny or denigrate a question that many business leaders are starting to ask themselves, absent solid cues from their respective governments – just how ARE we going to leave the world a better place?

Questions, questions

The questions raised remain. I am for instance concerned that despite all the Congressional face slapping at Facebook and Twitter, there is a hesitancy among legislators to attempt formulating regulation. The American ideology of the free market suggests that lawmakers will content themselves with having issued a stern warning in the expectation that players will do the 'right thing.' I find that hard to believe given the supremacy of the profit motive.
Tim O'Reilly presents an optimistic conclusion in a well-researched presentation about what happens next in an AI-driven world. He talks about the jobs Amazon has created but without juxtaposing that with his otherwise well-known views on income inequality. He talks about unintended consequences and the need to consider how those are fixed. It's a good starting point but his call to arms that the answer is down to 'us' - as in technologists - falls short. Well-meaning people are not enough.
Quite what the next step looks like it is very unclear. Is it conceivable that hainvg watched the failure of what we currently know as capitalism that we should review the principles of Fordism as a way to inform the future? I don't know but it is at least worth putting on the table.
For now, I leave the last word with Stephen Hawking who, at the recent Web Summit in Lisbon, said this:
We cannot know if we will be infinitely helped by AI, or ignored by it and side-lined, or conceivably destroyed by it.
A grey colored placeholder image