Today’s debate about artificial intelligence and the future of the workforce often centers on dire warnings about how robots are going to steal people’s jobs. Albeit anxiety-inducing, these conversations raise critical questions about the role of artificial intelligence and advanced robotics within the labor market -- and the number and quality of jobs for people in the future.
But as it turns out, it’s unlikely that most jobs will be taken over by robots or AI agents in their entirety. The McKinsey Global Institute estimates, for example, that about ⅓ of activities can be fully automated for some sixty percent of jobs. So, while some portion of the necessary skills may be automated, the remainder of the tasks required to perform the job may very well stay the same.
Consider the case of machine learning software that's now being developed to identify melanoma. Today, that sort of analysis is done primarily by M.D’s. But even if the technology is broadly adopted, dermatologists won't be replaced by AI. Rather, they will use software to inform their diagnosis, prior to developing a treatment plan -- and sharing the recommended course of action with their patient.
When forecasting the impact of AI, it’s important to keep in mind that today’s tech driven labor market shifts have been brewing for years. During the last 30 years, new tasks and job titles have accounted for almost half of U.S. job growth. Over the last decade alone, we’ve witnessed the emergence of new skills and roles -- from app developers to cloud computing engineers, user experience designers and data visualization experts -- that simply didn’t exist until this century. And the jobs of the future aren’t limited to high tech. LinkedIn’s Emerging Jobs report highlights “Barre instructor,” alongside machine learning engineer and data scientist, among job titles poised for growth.
A great example of this sort of market evolution is the story of Keith Broni, the first person in the world with the job title "Emoji Translator." His job is to help companies navigate safely through the emoji media minefield, a pictorial language that has exploded in popularity since emojis were invented in 1999.
And even when job titles have not changed, job functions and the skills required to carry them out have. In the 1980s, to become a bank teller was to enter one of the fastest-growing career spaces in the country. At the time, 531,000 people in the U.S. were employed as bank tellers. As ATMs and mobile apps threatened those jobs, the total number of tellers has declined, but it hasn’t disappeared altogether. Instead, their jobs have evolved. Today’s bank tellers look more like Apple Genius Bar employees; they help customers troubleshoot apps, walk customers through the how to’s of bank technology, and otherwise provide tech support.
Of course, there’s no getting around the fact that AI and automation will decrease demand for some kinds of jobs and their component skills. There is growing consensus that automation technologies could significantly reduce the demand for repetitive, predictable tasks in existing categories of jobs like drivers, telemarketers, fast food cooks and paralegals.
Workers in these and other fields will be faced with the urgency of embracing a shift fueled by a shrinking shelf-life for many skills. And the challenge will not be limited to blue-collar workers. A recent World Economic Forum report estimates that about 50 percent of the subject knowledge acquired during the first year of a four-year technical degree is outdated by the time a student graduates.
It’s a shift that will be felt as early as K-12 education. More than half of U.S. teachers now say computer science should be mandatory, but most schools do not offer any computer science classes at all. It’s not about teaching every student to code, but it is about the application of computer science to encourage computational thinking. It’s about giving students a vocabulary in computing, encouraging collaboration and communication, and helping them gain the confidence and ability to solve complex and abstract problems using technology.
Accelerated change in the labor market is also fueling a demand for better data and increased transparency to help employers find skilled workers -- and would-be employees identify the gaps between the skills they have and the skills they need.
Governments can play a big role in reducing labor market friction by utilizing more timely, actionable data, and moving beyond sometimes outdated workforce development systems. States like Utah are building public-private partnerships so that job-seekers can benefit from a more real-time understanding of where the market is headed.
Over time, better, more timely, and longitudinal data will better inform labor market debates with an understanding of how both public -- and personal -- investments in skill development can affect employment outcomes. Armed with better data and better training, we’ll be prepared to work alongside AI rather than get swallowed up by it.
Pablo Chavez is General Manager of Microsoft's US policy team, LinkedIn's global public policy team, and LinkedIn's Economic Graph research and civic engagement team.