This week, public-sector discussions centre around artificial intelligence and how technology will shape the jobs of the future. Research suggests future jobs will require people skills more than high-level engineering and maths, as most will work with technology, rather than build it.
There are two popular views about the impact of technological change in the workplace and the skills mix that will be required by tomorrow’s workforce.
One well-known fear is that various kinds of automation, powered by artificial intelligence and other emergent technologies will put masses of people out of jobs.
The other is the assumption that technological change means skills in science, technology, engineering and maths — particularly an understanding of computer programming, which is now widely seen as a vital part of school curricula — will be in increasingly high demand.
But analysis of labour force data from Australia and the United States shows that the big growth is occurring in “people skills” rather than STEM. And that makes sense in some ways when combined with the first fear. Logically, the more that automation takes processes out of human hands, the more our uniquely human skills come to the fore.
While plenty of people will find work in certain burgeoning fields like cybersecurity or the data sciences, and in the technology industry developing cutting-edge applications of AI, it is likely that the bulk of the workforce will be using more and more technology, rather than actually building it.
Not all jobs will be swept away by technology, some will be complemented by it, and in many cases those workers who are displaced will be able to transition to other jobs, perhaps at higher rates of pay, with a relatively small amount of extra training or education.
Estimates of how this will affect the labour market in terms of unemployment vary considerably, but the machines haven’t taken over yet. We still have our free will, and some experts are concerned about what we do with it.
“It is really important that we are not being passive about technology,” says Claire Mason, an expert on these issues from Data61, who worked on the aforementioned studies showing that the value of people skills, particularly around communication, is growing much more quickly than demand for traditional STEM skills.
Discourse analysis recently showed people commonly talk about technological disruption as “something that happens to us”, she adds.
“But the alternative is a much more active stance towards technology — that technology is something that humans have always wielded to shape and improve our lives,” Mason told The Mandarin.
“And that is what I think we need as a mindset, not that we all need to become coders, but that we all need to stay engaged with how we can be using the latest improvements in digital technology to do what we currently do most efficiently, in a new way, to add value, to enhance a customer’s experience.
“That is the mindset that Australia needs in order to maintain our position rather than either fear about technology, or this very narrow view about us all needing to become software engineers.”
Mason is speaking on a panel at the Australian Information Industry Association’s conference in Canberra focused on these very issues, among a long list of federal public servants and parliamentarians, academics and ICT industry leaders.
She thinks there is too much alarming rhetoric around, encouraging people in jobs like teaching or community services to feel “threatened and resistant to the technology” — but “it’s that attitude that will hurt us, rather than the technology itself” in her view.
Some of the research into skills trends showed relatively low growth in demand for traditional STEM areas but there is significant growth in the need for what Mason calls “generic” STEM skills, like using computers.
“So yes, in so many of the jobs of the future, people need to be able to work with tech. They need to be able to make sense of the numbers that might come out of a machine or a program, but they don’t actually need to program [computers], most people,” she explained.
Teaching coding in schools is sometimes said to be important for the same reason writing is important, even though few students will become professional writers; it teaches a basic understanding of how programming works.
Mason says that being able to use technology to “add value, to be more productive” at work will be important, across the workforce, but we don’t need a world of programmers.
A lot of workers will be able to learn how to use new technology that augments existing jobs, and others will be able to transition into new jobs that are created, which will likely maximise the value of the human touch.
Robots on the case
The Department of Human Services has begun implementing systems built around forms of AI — including as a decision-support tool that is relatively easy for staff to use.
Chief information officer Charles McHardie spoke about this on Tuesday at a forum about AI in the public sector, organised by the Institute of Public Administration Australia (full video below). He told his fellow public servants about the department’s exploration of AI products like IBM’s Watson, Microsoft’s Cortana and Google’s TensorFlow, and its three digital assistants.
Sam answers questions over the phone, Oliver’s job is guide people through the Centrelink claims process, and Roxy helps customer service staff when they have a tricky question about a claim they would otherwise have to ask a superior.
McHardie spoke about the jobs related to rolling out virtual helpers like Roxy — he said they needed to be “corrected and directed” repeatedly, like a child — but the research suggests these technologist roles won’t dominate the workforce.
Three questions rounded off an entertaining and thought-provoking keynote by Genevieve Bell, a cultural anthropologist and futurist who spent a long time in Silicon Valley and is now an Australian National University professor leading a new research institute that collaborates with Data61:
What metrics, beyond efficiency, will be necessary in the 21st Century workplace?
Where on the organisational chart will we put artificial intelligent technologies?
What does training/skill acquisition look like in a 21st Century learning environment?
The first concerns the possibility that standard ways to measure the value of technology might not be the best methods to judge the success or failure of AI-enabled tools.
“What are going to be the other metrics we might want here?” Bell asked. “If not efficiency, is it efficacy, because we’re talking about whole systems?
“Is it about engagement, because we’re changing the model of what that work might do? Is it also about other things that we are far less used to thinking about technology giving us, like the space to make things? Is it about creativity and, god forbid, that much-ballyhooed word, innovation?”
She said the continuing rise of AI was about more than the various technological strands that have come together, “… it’s also about ethics, and regulation, and reason. It’s about how do we think about the systems that will contain these systems.”
In Bell’s view, questions about security, transparency, trust, and “explicability” of automated systems have to be “part and parcel” of public policy and regulation as algorithms and AIs begin to play a larger role.
“This can’t just be about ‘it took us less time to make a decision’ — we might want to think about quality of decisions. There are all sorts of mechanisms we might want to involve here as we think about how we will want to use these things.”
The second question is just a humourous way of considering how AI fits into hierarchical organisations. Who can challenge or overturn automated advice or decisions? Will bots be equals, subordinates, bosses, or some of each?
Education and training up next for disruption
The third question calls out an issue that has occurred to Mason and her colleagues as well: standard education and training needs to change. Universities, Bell suggested, might be next to face digital disruption.
“There’s going to be a real pressure to deliver education somewhat more efficiently but also to allow people to dip in and dip out more, which is lifelong learning,” Mason told The Mandarin.
Another issue is that technological change is never evenly distributed, as Bell noted, repeating this oft-quoted observation of cyberpunk author William Gibson. In some places the future is already here, but in others, like Centrelink’s welfare payments platform, extremely old legacy technology is still in place.
This certainly affects the ICT skills DHS needs to employ; it has a need for software engineers that is hard to fill from the local market, but its massive seven-year project to replace the system might change that.
Theoretically, a huge chunk of current jobs could be automated, but there is reason to doubt how soon the potential will be realised, given old technology tends to linger, especially when it costs a lot to replace. And other jobs that make use of especially human skills will be created, Mason suggests.
The research into skills demand suggests workers won’t suddenly need to undertake “massive re-skilling” but rather small “tweaks” to their existing skillsets, she explains, through new “micro-credentials” for an increasingly digitised world.
Mason says ongoing data analysis could reveal useful details about the kinds of add-on skills people will need to keep pace with how their jobs evolve, or to switch over to new jobs that emerge, and Data61 has already begun helping the Department of Jobs and Small Business (formerly Employment) understand how it could do that.
Other speakers at the AIIA conference on the jobs of the future include the head of the ACT public service, Kathy Leigh, Attorney-General’s Department secretary Chris Moraitis, Digital Transformation agency boss Gavin Slater, Australian Public Service Commission first assistant commissioner Kerryn Vine-Camp and several APS deputy secretaries.
There will also be politicians like the opposition’s digital transformation lead Ed Husic, who has already thrown the cat among the pigeons with some comments published by Innovation Australia on Wednesday.