AI is going to change your career. IBM is showing how that can be a good thing.

Samantha Lee/Business Insider

• Workplace applications of AI in development include helping managers measure and run their teams effectively, and cost-savings through greater efficiencies.
• As disrpution roiled its sector, IBM used AI to help transition it to a skills-based company. Those tools help other companies now.
• This article includes an overview of AI in the workplace, including the top three trends to watch, plus an analysis of how IBM applied its innovations to solve its own problems.
• Read how AI is transforming health, transporation, consumer technology, and more in other articles from our special report, How AI is Changing Everything.

We're in the early stages of a massive technological transition that is spreading across every type of industry. Chief among the worries caused by this seismological shift is the loss of steady, long-term careers to automation - and the disappearance of the financial and emotional stability that jobs provide.

At first glance, the numbers seem to support this fear. Consulting firm Bain estimates millions of jobs will be lost in the US over the next two decades, displacing as much as 25% of the labor force. Ginni Rometty, CEO of tech giant IBM, has been even more direct with her predictions: she's said artificial intelligence will change "100% of jobs in the next five to 10 years."

But let's take a historical breather. After all, automation wiping out jobs isn't new; what's changed is the method: Deep machine learning is now capable of replacing more than just manual labor. This makes it easy to imagine robot bosses coldly monitoring our work and firing us when we don't meet pre-calculated criteria - HAL 9000 in a suit and tie.

And yet, the inevitability of AI's increased presence in the workplace doesn't have to cause despair. While Rometty is clear things will change, her company has been proving that when used with managers and employees in mind, machine learning can be a powerful asset in keeping jobs safe.

Over the past five years, IBM has developed a suite of AI-powered tools that help recruit and map career paths, and even determine salaries. And because the company is so influential, the programs they introduce could find their way to your office soon.

Major corporations are aware of the fear that often accompanies these changes (not to mention the bad press). Amazon, which has faced criticism for the way AI is used to terminate underperforming factory employees, announced in July 2019 that they would invest $700 million over the next six years to retrain 100,000 employees whose jobs are at risk of automation.

The professional services company Accenture has also been an industry leader during these changing times, investing about $1 billion in job training each year over the past four years. So far, it has retrained 300,000 workers. Its AI program (the whimsically named Job Buddy) helps employees whose roles are in danger of being replaced find new opportunities within the company. "What it does is it tells our people that, look, this percentage of your job is likely to be lost to automation; your skills are adjacent to these skills, so go take this training," Accenture's HR head Ellyn Shook told Business Insider.

Brands around the world are using AI to enhance their human-resources departments. Research shows that more than 70% of employers and recruiters use some form of automated résumé parser to process CVs, create summaries of the candidate pool, and even rank candidates. Consumer-goods giant Unilever and mobile provider Vodafone have bet that AI can help eliminate bias from the job-application process. HR-services firm Pymetrics develops quizzes that test entry-level job candidates for traits like focus, memory, and propensity for risk. Its algorithm determines whether a candidate is worthy of a closer look from a hiring manager.

Pymetrics CEO Frida Polli told Business Insider her company is not looking to replace hiring managers at large companies. Rather, they want to free them from having to consider thousands of applications when they could instead be focusing on the top candidates. "People are being elevated to a much more strategic role with the use of it," she said.

AI in the workplace is still in its early stages, and these experiments are largely confined to companies with access to advanced technology. But the tools are worth exploring not just to improve output and productivity, but to enhance employees' well-being.

Sven Hoppe/Getty Images

Top 3 opportunities for AI in the workplace

Making managers more effective: Nascent

A manager augmented by AI is vastly superior to a program that outright replaces a boss. At IBM, managers use an AI tool called Compensation Advisor that sorts through historical data on thousands of employees and competitors to present a salary range managers can work with, saving time and reducing the scope of judgment required.

Providing employees with career-advancement opportunities: Nascent

The developed world is at another inflection point, where technology is transforming jobs at a rapid rate across industries for all professionals. IBM's Blue Matching program analyzes data automatically collected about a worker's skill set and weighs that against requirements for other roles within the company. The employees can then use the Your Learning hub to acquire new skills, from project management to cybersecurity basics.

IBM

IBM's Your Learning platform creates a custom collection of skills-based lessons employees can use to advance their job.

Cutting time-consuming tasks to boost productivity: Growing

Automation has always been used for tasks that a machine could perform more efficiently than a human. However, a new generation of tools is not only cutting time, it's enabling managers and employees to work more effectively. IBM employees used to have to answer long surveys about their skill sets, then go over them with their boss. Now this data is collected by AI, giving teams more time to focus on the discussions rather than the paperwork.

How IBM used AI to transition to a skills-based company

In 2014, IBM's leaders realized the tech giant was facing massive disruption. Cloud computing (software distributed online instead of through hardware), blockchain (secure ledger tech), and advancements in AI were redefining their industry. Left unchecked, the drive to automation would leave much of IBM's 350,000 employees without a job.

Diane Gherson, who'd been named IBM's head of HR the previous year, knew that if the company was going to thrive in this era of change, it had to shift into becoming a "skills-based organization" - that is, employees would now be assessed primarily by their collection of skills, allowing them more flexibility in a rapidly changing environment. And AI, which was partially responsible for this sudden and drastic shift, was going to help them adapt. "It was a huge leap," Gherson told Business Insider.

Veteran IBM researcher Anshul Gupta was one of the leads in developing AI-powered tools for this purpose. Fundamental to his team's mission was finding solutions to problems that already existed, rather than creating software they found interesting and seeing what stuck. First on his list was figuring out how developers would approach the problem. He told Business Insider that a question he had to ask himself was "How do you think about enhancing their experience through your offerings, but in their workflow?"

As Gupta and his team set about crafting those tools, another powerful AI was being developed elsewhere at IBM. The company introduced in 2016 a consumer version of Watson, the question-answering AI that famously made its public debut as a "Jeopardy!" contestant. Kelli Jordan, who's overseen IBM's New Collar Job training program and its apprenticeship initiatives, asked developers to find ways to implement Watson into their own workplace.

"When you start to think about careers and development, especially at a company as large as IBM (and especially in the environment that we're in now where skills are changing so rapidly), it's just hard for an employee to really take it all in and make the right decisions and understand where to look," Jordan told Business Insider.

The resulting suite of tools that branched out of Gupta's research and the Watson initiative touches nearly every aspect of the working experience, from answering a job candidate's questions to alerting managers when employees should be considered for promotion - or when they're in danger of missing a quota.

IBM

IBM's Your Learning platform creates a custom collection of skills-based lessons employees can use to advance their job.

IBM has determined through quarterly assessments of financial and performance impacts that these tools saved the HR department $107 million and "thousands of hours" in 2017. The company has since greenlit further development on the tools.

Of course, as with any AI, each program has to learn through data accumulation and recognition of patterns. So when the skills assessor first rolled out to an initial batch of employees around the world, it didn't have data enough to work with.

Peter Usacov, an IBM project manager based in Hungary, volunteered to be an early adopter of the skills assessor program. He told Business Insider that he and his manager found it useless at first, though he decided to keep engaging it to see what happened. "It was like, 'Oh, this is rusty and doesn't really give me anything,'" Usacov said. "Two, three months later, I'm like, 'Oh my God, what happened in between that this became so smart and accurate?'" Usacov said he is now using it for help as he considers new career paths at IBM.

The company claims that its skills assessor currently has between 85% and 95% accuracy.

IBM also employs hundreds of AI experts to fine tune the various tools' data sets, to adjust for outliers or decisions stemming from unwanted human biases. Gupta told Business Insider that the compensation advisor is proof of how confident he and his team are in the accuracy of their programs' ability to gather and analyze vast amounts of data, since such a tool is not something you would undertake and use if you were "just playing around" with AI. So far, it's been used for 2,000 pay recommendations.

The world where AI is capable of matching or exceeding human intelligence is "very far away," Gupta said. Still, the tech is now at a stage where it can be trained to do a particular function, like determining how much to consider paying an employee "very, very well."

"That's an area where you enhance the employee experience, you deliver good business outcomes, but you also transform the roles of the team itself," he added. "I think that's a practical reality of what we've seen as we've deployed it within IBM."

The Takeaway

Hollis Johnson/Business Insider

"One-hundred percent of jobs are going to change with artificial intelligence - we know that, right? But that doesn't have to be terrifying. ... If you can make people educated about how to make change and engage them in it, then you're going to end up in a better place. And people are going to be happier because they're doing more value-added work."

- Diane Gherson, SVP of Human Resources at IBM

Get the latest IBM stock price here.