Everywhere you turn, there’s another news story emphasising the impact artificial intelligence will have on the enterprise this year. Along with that are headlines screaming about Silicon Valley giants siphoning off talent at a time when machine learning has become a huge driver in the battle for digital transformation dominance.
So, if you’re looking to make good on the promise of AI, where can you turn for talent? Desperate times, it is said, call for desperate measures. Many organizations are dealing with the AI talent shortage by forming partnerships with universities and by training and building from within.
If you think this is all a lot of hype, consider that by 2030, the global GDP could be up to 14 percent higher, or $15.7 trillion as a result of AI, making it the biggest commercial opportunity in today’s economy, according to the recent PwC report “Sizing the Prize.”
“If your business is operating in one of the sectors or economies that is gearing up for fast adoption of AI, you’ll have to move quickly if you want to capitalise on the openings, and ensure your business doesn’t lose out to faster-moving and more cost-efficient competitors,’’ the report advises.
Regardless of whether a company is in a sector where the disruptive potential is lower and adoption is likely to be slower, PwC says, “no sector or business is in any way immune from the impact of AI. … The big question is how to secure the talent, technology and access to data to make the most of this opportunity.”
This is something CIOs will be grappling with as more organizations deploy AI-oriented initiatives in the next couple of years. Staffing skills is the No. 1 challenge for 54 percent of CIOs looking to adopt AI, according to Gartner, which has deemed 2018 as the year “AI Democratization” begins. But some CIOs are facing a double-whammy as 37 percent are still struggling to define an AI strategy, the research firm says.
“The challenge of creating an AI strategic development plan parallels the staffing challenge, as having AI-savvy workers and executives benefit organizations actively working to set strategy,” according to the Gartner “Predicts 2018: Artificial Intelligence” report. That said, the firm believes by 2020, 85 percent of CIOs will be piloting enterprise AI projects using a combination of buy, build and outsource efforts.
Consequently, management skills need to evolve, since managers have “only recently come to understand and rely upon advanced statistical techniques that extract ‘signals from noise’ to improve decision making,” the Gartner report finds. “This is the starting point for managing AI-based systems and services.”
Jobs will change, but people will still be needed
Specifically, demand is growing for data scientists, robotics and AI engineers, and workers with experience in deep neural networks, big data and analytics, among other tech specialists.
But given the pace of technological change, increasing computational processing power and a lower barrier to entry, these workers are hot commodities and have become elusive.
There is a broad assumption that AI will eliminate jobs and tasks, and while some will be consumed by machines, human workers will not be replaced, but they will need to master new skills to work alongside AIto deliver value that only humans can provide, experts say.
“Companies that become overly reliant on machines to complete work without devoting time and resources toward building human-driven skills could face a significant AI staffing and skills shortage,’’ says Ben Pring, vice president and director of Cognizant’s Center for the Future of Work.
Organisations should invest in building the following employee skills for the new workplace landscape: analytical thinking; verbal & written communication; design; decision-making; interpersonal skills; and global operating, among others, according to Pring.
Building from within
Finding external AI talent is very difficult, says Greg Layok, senior director of Chicago-based tech consultancy firm West Monroe Partners. “There’s very few unicorns out there and it’s very hard to find people with all the capabilities you need.”
In the past five years West Monroe has interviewed lots of people, he says, “and it’s hard to find ones with an academic background, specifically in mathematics or statistics,” who also have an understanding of the tech tools used to build machine learning systems in corporate environments. The ideal candidate also has solid communication and soft skills. “Those people that have all those capabilities … are worth their weight in gold right now.”
Out of “hundreds of individuals” West Monroe has interviewed for AI-related positions, “we’ve found one or two that have those full set of experiences and capabilities,’’ Layok says.
West Monroe’s strategy has been to hire people who are “highly collaborative,” meaning a willingness to work on a team and do work outside of their discipline on things that might be foreign to them. “What’s interesting is, there are definitely programmes, master’s programmes, where we’re getting people with good hard skills, but they don’t necessarily have the business acumen to apply their data science skills against a business problem,” he notes. So West Monroe focuses on trying to “blend people together and take the best of every individual.”
Recent college graduates who have aptitude and are open to their learning environment are great prospects, Layok says. “We pride ourselves on building the next generation of leaders and believe we can grow AI experts from within.”
That sentiment is also shared by Dr. Ben Waber, a visiting scientist at the MIT Media Lab. “Learning on the job is still cheaper than spending $40,000 on recruiters to hire the same person — and then you have to onboard them anyway,’’ he says. But organizations still need experienced people who understand the deeper complexity of an algorithm, and will have to pay top dollar for them, he adds.
But thinking that “every single person has to have a PhD and five years of industry experience, you’re just overthinking this,” says Waber, who is also president and CEO of behavioral startup Humanyze.
The company is hiring seven to eight people per month and expects to have close to 100 employees by the end of the year. Like Layok, he says that Humanyze looks for people with strong mathematical knowledge, and ideally, statistics. “That’s better than having a programmer,” Waber says, “because fundamentally, the state of the art in AI learning are straightforward statistical tools.”
Establishing an AI talent pipeline
“We view the talent shortage in data science similar to how professional sports leagues have worked their way back in recruitment: We use talent at earlier levels,’’ says Mark Clerkin, a data scientist at venture capital firm High Alpha, in Indianapolis, Ind. “We have relationships with universities and engage in learning projects and doing speaking to have access to talent at that level before [graduates] are placed.”
High Alpha also gives students still in school “real-world experience, and we give them meaningful projects and get to know each other, so it’s kind of a long running interview.” That essentially gives the firm a talent pool to draw from, he says.
The firm also communicates internally about what AI is and what can be done using machine learning technology. “The goal there is to demystify the aspects that go into machine learning,’’ Clerkin says. “By doing so, you start to surface people internally who are interested in doing AI components,” such as modeling and extracting data from files and massaging it. Staff with some engineering or math skills make ideal candidates for this type of work, he says.
The AI talent shortage is real
If an organisation is looking to build a scalable AI system, it will also need people with “pretty sophisticated back-end experience,’’ he says.
“We’re certainly deep in the hype cycle [of hiring] right now,’’ says Waber. “Everyone and their brother is doing AI right now.” At the same time, he says, “some companies are paying lip service [to AI] but not doing hiring.” In general, though, there is strong recognition that having AI skillsets is important, and organizations are gearing up to both hire and train from within. This will remain constant for a while, Waber believes.
“Gartner uses the term ‘citizen data scientists;’ people in your organization that you bring to the other side,’’ says Layok. “I think university programmes are getting better, but my experience is they’re not there yet. Graduates are getting better and better every year, but that alone won’t satisfy the demand” for AI workers. If companies haven’t looked at internal development yet, it’s probably a good idea to start now, before their AI efforts ramp up.
Universities are doing their part, and many have begun offering meatier undergraduate, graduate, and professional certificate programmes in AI and machine learning. Yet even with schools now producing many graduates with valuable deep-learning skills, Gartner says few of them have the intuition that delivers great foundations for a successful deep neural network model.
Technical skills — especially for deep learning — remain limited and are still evolving, the Gartner report notes. “We still do not understand how to reliably configure a DNN [deep neural network] to deliver useful results, and the long turnaround time on DNN training makes for a long evaluation cycle.”
However you look at the AI talent shortage, it’s not an either or proposition but an “and,” notes Layok. “You need people with an aptitude for this kind of work both inside already, and from recent graduates with data science degrees.”
But he’s not suggesting it will be an easy process. “Building this talent inhouse has been a huge challenge for us in the last four to five years. I finally feel like we’re over the hump, but we’re not done.”