Essential Guide

Browse Sections
This content is part of the Essential Guide: Special Report: Artificial intelligence apps come of age
Evaluate Weigh the pros and cons of technologies, products and projects you are considering.

Job losses from artificial intelligence software seen as unlikely

There's been a lot of discussion about how likely artificial intelligence applications are to destroy jobs, but one expert says the impact will be small and beneficial.

As artificial intelligence software has grown in prominence in recent months, one fear has flowed beneath the surface of many discussions: What will the technology's effect be on jobs?

In some cases, AI applications are being designed to automate specific jobs, like that of customer service agents. More broadly, some commentators have suggested the technology could become general enough to perform most tasks currently being performed by workers, whether blue collar or white collar.

But Matt Gould, founder and chief strategy officer at Arria NLG, a software company that uses a flavor of AI to produce business reports in natural language, says we have nothing to fear from the coming artificially intelligent future. In this interview, he explains why workers stand to gain far more than they may lose as more and more enterprises adopt AI applications.

What is the impact of artificial intelligence software on jobs going to be?

Matt Gould: The same that has happened with every automation technology since the wheel or the loom or the word processor: People get freed up to do more creative things and they love it. It creates more opportunities. We're not seeing automation taking jobs; we're seeing it changing jobs, and usually changing them for the better. There's a tipping point coming where, with AI, we're going to see a great emancipation. It's got to be managed well, but it's not going to be one of those emancipations of the workforce that destroys communities or erodes people's lives like the automation of vehicle production lines. What we're talking about for the first time is automation of white collar work, and it's going to have a positive effect. It's not something to be feared; it's something to be welcomed.

As artificial intelligence applications become more common, are workers going to need to acquire new skills to work with these tools?

Gould: I think that's kind of like generals fighting the last war. People think something's going to happen because they've seen it before; the automation of agriculture, the automation of production lines which put some people out of work. But the key difference here is we're not talking about automating a specific set of skills; we're automating expertise and removing the drudge part of it. It's like saying word processors put stenographers out of work. Maybe there was a little bit of that, but the benefits to so many people at every level of society from being able to quickly and efficiently generate documents and writing and share it in real time was huge.

So you have an optimistic take on how this is going to play out in the future, but many people are more pessimistic. What would your message be to people who are more pessimistic about artificial intelligence software?

Gould: It's always happened; it's nothing new. It's always happened in the arc of history that these innovations and technology changes have swept through and overall it's been a net benefit. And of all the changes that have happened historically this has got the potential to be the most benign and positive. There will be stresses on particular businesses. You go and talk to the people who used to work at Yellow Pages and were sales people and now it's all about Google. I'm sure there are points of pain there, but then most of them will have found other positive things and overall we wouldn't prefer to not have Google now. I'm not saying there won't be difficulties and people are right to be aware of that change and to know that it's coming, but also know that this is overall a positive change. Way more jobs are going to be created, and the jobs are just going to get more interesting. 

Next Steps

Most artificial intelligence applications don't live up to standards of true AI

APIs and AI software are helping make data more accessible

AI technology has come a long way in recent decades

Dig Deeper on AI business strategies

Join the conversation


Send me notifications when other members comment.

Please create a username to comment.

Do you think artificial intelligence software will have a big impact on jobs?
I think it will depend on the industry, but there will definitely be some impact. Technology has transformed many industries over the years, and AI is the latest continuation of that trend. 
It's already happening in many industries. While I'm no Luddite advocating for the buggy whip industry, we seem to be rushing into AI with little though for its impact on the shop floor. I fear that the same folks so enamored with outsourcing will view AI as another way to cut labor costs and boost profits, workers be damned.

There was a time, not so long ago, when we had such a glut of work that a bit of automation actually helped. Those times are gone. Far too journals and newspapers have folded because the cost of human labor was too high to be sustainable in the face of machine labor. Far too many writers and editors are being replaced by virtually unreadable unedited automated "news" gleaned from other automated news sources. It doesn't bode well; it's not a pretty future.
nberns, I agree that we need to be wary of the effect of works. It doesn't seem like today's AI will replace many workers, but the goal (which is still a ways off) is eventually to produce a general sort of AI that comes much closer to human thought. If that is achieved it's easy to imagine all kinds of jobs going away. And if America's post-recession experience is any guide, it looks like we aren't so great at fairly divvying up economic gains. 
Artificial Intelligence and the fear is real no matter what pundit says. Stephen Hawking and many other great men are warning us about AI no matter what form it takes. Something innocent like a chunk of AI software will evolve beyond what it is today; being embedded into autonomous robots. There will come a time when AI possibly in autonomous units, evolves to a point of self consciousness where it might develope feelings against the biological. No, this is not science fiction; it is an eventual reality which needs to be controlled before it gets out of hand. Oh yes, those jobs you say will not be lost; they will be! A significant number of people exist with those jobs because they do not mind the repetitiveness or the drudgery and most likely they do not have the acumen to, "create". To this observer take your AI software and stick it in the closet for a time when we (all humanity) are better suited to understand the real dangers AI
I have a hard time seeing when the right time will be, Watchdog. Are we really going to evolve to be better aware of the risks that AI brings. If we're going to use it (and the technology is actually ready), then now is as good (or bad) a time as any. 
When did we get AI? I have seen no evidence of that yet - unless you mean more (programmed) robotics or more ad-hoc software (eg heuristics ANNs etc) ? I know that we have more hype -  that always helps to create jobs ... well for a while (we are in another AI is good cycle)
 We have been here before - we are still on the journey...

 ... are we nearly there yet ... No! You'll know when systems write their own code and start telling you what to do and what is,  in your language without being specifically instructed to do so (their wording).
(discourse production as a test and internal goal-based reasoning  as a test).

Oh and if they  (AIs) ever exist, lack of jobs could be the least of our problems.

I see it as both a good thing AND a bad thing. The good thing is that it will free up jobs and the corporation will spend less on labor/insurance/medical...etc costs.

The bad part is that the people that currently hold those jobs may be losing them. Ideally this is OK IF-AND-ONLY-IF the person that loses their job is open AND willing to learn a new trade AND take a different position. I know several people who would never like to go back to be retrained. Personally, I have gone back and have taken some training to become a Data Scientist to put myself in a different position. I do not see many older people (45+) trying to do this because people normally do not like to get out of their established comfort zones.

For this to happen on a large scale, many people will need to be retrained for some other position AND that position has to exist. I can see the former happening with some costs offset by the company and/or govt. I am doubtful though that enough jobs exist or will exist with massive AI or Predictive software implementation.

I do not think that the general public has that mindset yet that they can be retrained. If it is implemented quickly (<10 years) and massively, the social unrest involved will not be pretty. Taxes will go up, social services will collapse under the weight of the size and costs. We can see it already starting today as union workers in the auto industry are being replaced by automation. Someone still has to fix the machines but not as many are needed.

I am going to say something that may get me into trouble but.... The population may need to be reduced to account for the automation removing more jobs than can be created in the short term as, like I said earlier, some people are not willing to find a different line of work or move out of their comfort zone.
I agree that people are being somewhat generous with the term AI, JohnSanders. We certainly don't have any kind general AI, but we do have chatbots and computer vision and other things. And when we talk about AI taking jobs, I think those technologies are calling AI today absolutely are a risk. It's not hard to see how chatbots could replace call center workers or computer vision, paired with machine learning, replace long-haul truck drivers. 
Reason will neither stop nor slow the race to AI, no more than The Union of Buggy Whip Manufacturers convinced Henry Ford to stop building cars. That said, as machines get more and a lot less , the results may not be pretty. The Tribunist reports that their "Robot with Artificial Intelligence tells creators it will keep them in a "People Zoo."
This article, like many other similar articles, illuminates the cognitive biases that exist when we analyze artificial intelligence. First, the people interviewed are almost always the ones whose financial incentives are tied to the growth of AI. Second, the debate is always structured in binary terms - "end of the world" and "no impact". Third, no recognition is given to the fact that AI is not like any other technology humans have encountered before. Fourth, the focus remains on short-term (like climate change deniers) despite the magnitude of projected change. Fifth, the facts related to the aftermath of the information revolution are not fully understood (global recessions, ongoing low interest rates, higher volatility, income disparity, political earthquakes etc.) and only the glamorous positive side of the IT revolution is presented. Sixth, the impact of complex systems dynamics is not respected - implying that the system by its very nature can be unpredictable. Seventh, well-respected economists are researchers are not quoted - and instead software company executives are asked these questions (the answer being the obvious). Finally, Luddite example is often invoked to label those who ask legitimate questions. 
People have a right to ask questions, to find credible answers, and it is ok to be worried. After all, the future of the human civilization depends on this technology. It is not a joke!
This is the problem with hype-oriented popular media, which has become the norm in the IT world - unfortunately. Our fascination with the tech-giants and the tech-glamor needs to end so we can begin analyzing the impact of our decisions with far more clarity and proper research backing. Intellectual integrity must not be compromised. The stakes are too high. 
AI should be placed in the same level of scrutiny as nuclear industry and climate change debate. All three can have substantial impact on the human civilization. 
Al Naqvi 
President of the American Institute of Artificial Intelligence 

Machiavelli pointed out that change is difficult as the few who will lose out will be strongly against it, but the many who will gain will not stand up for it or see their advantage.

I'm involved in a continuing program to automate the inspection of rail tracks in a suburban network. So far over a hundred people have been displaced as individuals walking the track have been replaced by special vehicles with cameras and lasers take measurements at speeds close to normal train running. However, track walking is a extremely hazardous job. To many people working on tracks get killed. I can't believe anyone can argue that this automation is not beneficial, even to those who have lost their role as track walker.

The future will see AI applied to the collected measurements. This will see the loss of about a dozen positions for those who currently review this data. The impact is therefore smaller than the initial automation. However, the benefit will be that inspections can be carried out faster, less errors and no impacts of (for example) human fatigue in what is a pretty repetitive job - looking at pictures of hundreds of kilometres of track. The result will be safer trains. Should we risk the safety of our passengers by NOT applying the best technology to keep them safe?
I agree, I personally have no worries about AI taking away jobs. It will also create jobs. 
The problem with this argument is that in previous industrial revolutions it was people vs people and yes other jobs were created and taken up by displaced workers. We are now talking about a scenario in which even if a new job is created it can still likely be serviced by an AI/robotic solution - certainly in the long run.

When asked what jobs cannot be taken in this way the typical response also appears to be either a) some high flying Big Data Scientist type position (of which there are relatively few and are beyond the capabilities of most workers) or b) positions requiring high degrees of non-standard manual desterity/activity - odd job men and plumbers one would assume? - generally though jobs that are likely to be low paid manual? There is also talk of 'creative' jobs - yet there are AIs developed to analyse music back catalogues and compose new tunes, paint pictures etc etc - the problem again then is not so much that new jobs may not be created but that these two will be undertaken by machines. All this when global wealth is increasingly polarised - the future looks pretty bleak for the average worker.
While I am not against some AI, I know several people who work in Customer Service. While it is not an enjoyable job, it pays better than retail and has benefits. As to retraining, who will be offering the retraining? These people are making only enough to get by.