There are many differing opinions on the impact of artificial intelligence (AI) on our worklives, from dazzling to dystopian.
Bill Gates, for one, sees the bright side of things. "Think of all the time we spend manually organizing and performing mundane activities, from scheduling meetings to paying the bills," he writes in the foreword to Satya Nadella's new book, Hit Refresh: The Quest to Rediscover Microsoft's Soul and Imagine a Better Future for Everyone
, "In the future, an AI agent will know that you are at work and have ten minutes free, and then help you accomplish something that is high on your to-do list. AI is on the verge of making our lives more productive and creative."
The dark view of AI is taken up by tech visionaries such as Elon Musk and Kai-Fu Lee, founder of venture capital firm Sinovation Ventures, who warns
that artificial intelligence will replace up to 50% of all jobs within the next few years.
So does AI enrich or take away jobs? The answer probably lies somewhere in between. In the meantime, there's probably more confusion than clarity. I was at a session at the recent Strata data analytics conference in which the speaker, Edd Wilder-James
, asked if members of the audience could provide a definition of artificial intelligence. No one offered to provide one, underscoring the speaker's point that it is a vast, unexplored area with ramifications that have yet to be understood. Wilder-James offered up his own definition, by the way: "AI is an extension of taking some of the things we analyze with big data and putting some decision making framework on top of that."
Spiros Margaris, a venture capitalist and thought leader in the digital financial services space, recently provided
a sharp definition of AI, and what to expect of it:
"Think of AI as computer intelligence, or in other words, machines that try to simulate (or one day to exceed) human intelligence and perform tasks cheaper, faster, and better than humans. AI has been around for a long time, but it is the explosion of data, computing, and analytics power that has brought the new opportunities - previously envisioned only by scientists and science fiction writers - that are now increasingly becoming realities."
It seems many people these days tend to agree more with Margaris' and Gates' assessment of the possibilities of AI, as confirmed in a recent survey of 650 employees released by CCS Insight. Currently, only 13% say they have been directed affected by AI, but an additional 40% expect to feel its impact within the next three years. About half of the survey participants expect AI to increase the number of jobs available, and 26% expect it to alleviate them of mundane day-to-day tasks in their jobs. "Employees also expect gains to include improved performance of equipment and machinery, the utility of assistive features in productivity and collaboration applications such as automated e-mail replies, intelligent scheduling and calendar features, and quick access to documents and search features," the survey report's authors add.
So, how should people prepare for the emergence of artificial intelligence, and how will it impact jobs? There will be two effects: it will change the nature of many jobs and professions, threatening some and wringing out inefficiencies in others; and it will create new opportunities not only in programming and development, but also in understanding the power and pitfalls of AI, and organizing AI resources into entrepreneurial ventures that meet market needs.
Some of these aspects were explored in a series of interviews conducted by McKinsey & Company at the Aspen Ideas Festival in June. As Andrew Ng, cofounder of Coursera put it, "We now see a surprisingly clear path for AI to also transform every single major industry. Everything ranging from much better healthcare to more personalized education to much more efficient retail and manufacturing to self-driving cars. This will displace a lot of jobs, everything ranging from call-center operators to, when self-driving cars come, the millions of truck drivers and maybe taxi drivers whose jobs will be affected. But this is true for white-collar and blue-collar workers."
Even medical professionals need to look at the long-term implications to their jobs, Ng adds. "AI's getting really good at reading radiology images," he says. "If any of you have a son or daughter or a friend graduating from medical school with a radiology degree, I think they might have a perfectly good five-year career in radiology, maybe even 10 years. But I wouldn't plan for a 40-year career doing that same radiology job today."
Yes, many types of tasks may be subsumed by AI. AI will affect just about every job in some way, whether or not they are technical positions. While this will not appear in most job descriptions, organizations are hungry for people who can take a high-level view of analytical systems. Marketing and customer service managers, for example, will need to understand AI enough to be able to design applications and interfaces that enable automated chats.
There's another high-level skill that will also be needed. A point raised in the Aspen discussions was the problematic and inherent biases that may be embedded in AI, and thus, too much trust being placed in machines with flawed, or even dangerous, logic. The ability to design inherent biases out of AI systems, and to provide critical thinking over and above automated decisions, is is a skill area that needs to be developed and brought into every organization considering AI.
"With the rise of automation, you definitely have conversations about jobs that are going to be lost," says Joy Buolamwini, founder of the Algorithmic Justice League at MIT Media Lab. "But I think something we're not talking as much about is who then become the gatekeepers for the jobs that are there? Even now, you have automated systems going through applications for jobs, looking for specific patterns. Those specific patterns might reflect prejudice in selection from prior decision makers. So now what you do is you embed that prejudice, potentially, if you're not intentional about checking for bias or trying to take measures to ensure fairness. Data is destiny. If you have biased data, you're destined to have bias in your outcomes or your predictions if it's left unchecked."
It's very important, in fact, that business leaders to develop the skills and get involved in designing AI and machine learning algorithms, as these can quickly get away from businesses, with decisioning wrapped up in unknown logic. At the same time, there is insatiable demand for "storytelling" skills -- often seen as part of data science, but extending well into other jobs -- which can deliver tangible scenarios to decision-makers on what AI is capable of delivering to the business.