When I edited our new Trend Analysis Report on artificial intelligence (AI), AI: The time is now, one of the findings troubled me: A large majority of our 302 survey respondents from communications service providers (CSPs) and their suppliers ranked concerns about AI displacing staff last on the list of potential challenges – indeed, only a third of CSPs cited it in their top three challenges.
It’s true that it’s very early days for AI in the telecom industry, which was another of the report’s major findings, but I still expected CSPs to be thinking more about the impact the technology is going to have on jobs and people. It will be profound because so many processes and operations are still performed manually. In the internet of everything manual processes simply cannot support the volume and velocity of changes required, so telcos are going to have to adopt virtualization, cloud, automation, and AI, including machine learning. They have no choice.
This leaves me wondering: Are CSPs drinking the Kool-Aid that so many technology companies are peddling when they say that AI will create jobs rather than replace them, and that people will find work alongside machines? Or are they in denial that automaton is coming for the jobs of many customer service agents, field service technicians, network engineers and network managers? Or worse, do they just not care what happens to the third of employees they’re going to cut?
About half of jobs can be done by robots
Late last month, McKinsey & Company released a new report on the future of work called Jobs lost, jobs gained: Workforce transitions in a time of automation. The report follows one released in January, which found that globally about half the activities people are paid to perform could be automated using current technologies. While McKinsey doesn’t single out telecom in the latest edition of its report, some of the findings are worrisome:
- 60 percent of occupations have at least 30 percent of constituent work activities that could be automated using current technology, but the proportion of work that will actually be displaced by 2030 is likely to be lower because of technical, economic and social factors that affect adoption. Depending on the country, somewhere between almost zero and one third of work activities could be displaced by 2030, with a midpoint of 15 percent.
This might sound like a reason to be optimistic about jobs, but even at 15 percent there are hundreds of thousands of telco employees worldwide whose jobs are at risk, and not all of them will be eligible for retraining or will want to retrain. Back in 2016, AT&T Chairman and CEO Randall Stephenson told The New York Times that the company expects employees to retrain (at their own expense) or lose their jobs. AT&T doesn’t really have to worry about the folks who opt to “ride the copper train all the way down” as one employee put it, because company executives expect to be able to do without about 30 percent of the workforce of 256,000.
We had difficulty getting CSPs to cite real numbers in projecting job losses to AI and automation, but one respondent whose company has set up an automation program within operations said that in the next four-to-five years his company expects to be able to reduce operations staff by more than 30 percent.
“The more systems you have which eliminate the human factor, the more instabilities you remove,” he said.
An even more sobering finding from the McKinsey report states:
- Even if there is enough work to ensure full employment by 2030, major transitions lie ahead that could match or even exceed the scale of historical shifts out of agriculture and manufacturing. By 2030, 75 million to 375 million workers (3 to 14 percent of the global workforce) will need to switch occupational categories, and all workers will have to adapt to work with machines.
The argument I hear most often from people who try to quell fears about job losses is that this technological revolution will be no different from previous revolutions when humans learned to adapt (albeit painfully) and found new work to do despite the job losses to technology. McKinsey seems open to this argument as well, but nevertheless offers these words of caution:
“On many dimensions, we find similarities between the scope and effects of automation today compared to earlier waves of technology disruption, going back to the Industrial Revolution,” the report states. “However, automation going forward might prove to be more disruptive than in recent decades – and on par with the most rapid changes in the past – in two ways. First, if technological advances continue apace and are adopted rapidly, the rate of worker displacement could be faster. Secondly, if many sectors adopt automation simultaneously, the percentage of the workforce affected by it could be higher.”
This is not normal disruption
I’d argue that the likelihood of both of those things happening is very high and that AI is in fact a different kind of disruption. Industrial revolutions in the past involved advancements in power and speed but still required human intervention for perception and decision making. AI obviates humans altogether.
In the 29 years I’ve been writing about telecom, I’ve never seen a technology more disruptive than AI. It’s true that operators are moving slowly to adopt it along with virtualization and automation, but they will do so, which means that the time to plan for the impact is now.
“We should embrace these technologies but also address the workforce transitions and challenges they bring,” McKinsey urges. “In many countries, this may require an initiative on the scale of the Marshall Plan, involving sustained investment, new training models, programs to ease worker transitions, income support, and collaboration between the public and private sectors.”
McKinsey suggests societies will need to address four key areas:
- Maintaining robust economic growth to support job creation
- Scaling and reimagining job retraining and workforce skills development
- Improving business and labor-market dynamism, including mobility
- Providing income and transition support to workers
- Investing in reskilling employees
- Supporting mobility and job rotation
- Targeting female talent
- Collaborating for skills with educational institutions and companies within and outside specific industries
We need to do more
Don’t get me wrong: I’m excited about the potential for AI and automation in telecom networks – I love self-service and I’m dying for an autonomous car. But I believe the AI revolution could have disastrous consequences if we don’t plan for its effect on people. We can’t just assume it will all work out fine.
During the coming year, the Forum will be spending a lot of time on automation and AI, and I hope this will include discussions about our responsibility as an industry. Specifically, how can we minimize the negative and maximize the positive impacts AI and automation will have on workers? Telcos requiring employees to retrain at their own expense is probably a little less that what’s necessary to minimize the negative, and it’s a million miles from the Marshall Plan-like scale McKinsey suggests.
It is perhaps the immense scale that puts CSPs off addressing the big picture. Where do you start? Is it possible for a single company to develop a plan to manage the impact of AI on its employees, or should it be addressed collectively?
New skills are crucial
AI requires new skills that must be developed at incredible speed, not something every CSP should or even could do independently. However, the WEF survey found that only 10 percent of ICT companies said they had plans to collaborate for skills within or across industries.
This is an area where TM Forum could have an impact. Our collaboration community could be an excellent place to figure out how sharing skills might work. What are the best practices for sharing talent among the members of a digital ecosystem? What kinds of business processes and agreements are necessary to facilitate sharing while protecting individual members’ interests?
I would love to see a Catalyst proof-of-concept project attempt to answer questions like these. If you’re interested in participating in a such a project, please contact me at [email protected].
We will also be surveying CSPs and their suppliers again next year about AI. Maybe then the impact on people will rank higher on the list of challenges we need to address.