“Microsoft lays off journalists to replace them with AI” was a recent headline which caught my attention. Not only because I am interested in the ethical issues of artificial intelligence, but also because I worked at Microsoft’s MSN News at its founding 25 years ago (it also spawned MSNBC, hence the “MSN” in the name).
We recognize that advancing technology displaces workers, as it has for over a century. We often think of these as lower-skilled workers: armies of ditch-diggers replaced by backhoes or the call center agents many companies would love to replace with chatbots. But we forget the thousands of salespeople superseded by web-based AI, or the travel agents and stockbrokers who were gradually displaced as consumers got direct access to data.
In fact, the word “computer” used to refer to humans who did calculations. Before 1950, masses of “computers” used pencil and paper to create invoices, total monthly statements and compute loan interest. All those “computers” have been replaced by technological computers.
Is the same thing likely to happen with AI?
At least for now, the answer is no. For the most part, any fears of AI displacing thousands of workers are misguided, in part because AI doesn’t currently have nearly the power that many ascribe to it. While in some cases AI is allowing us to work smarter and more efficiently, humans are still the ones making decisions on how AI is used.
AI, in fact, is a misnomer. Today’s AI has no “intelligence” because it has no common sense. Similarly, another term often paired with AI, deep learning, has no real “learning” because there is no understanding. We would be better served renaming these as the powerful, sophisticated statistical processes they are.
Related Article: AI Transparency and the Emperor's New Clothes
AI in the Next 10 Years: Bigger, Faster, Cheaper
What will the future bring for AI? Barring any major shifts in our thinking about AI, in the coming decade businesses are likely to see more of the same, just bigger, faster and cheaper.
As AI's capabilities expand, it can gradually replace more highly-skilled workers. Consider an employee making $25 an hour handling a transaction every few minutes. The minuscule per-transaction cost of a website begs the question, “Can AI produce equivalent (read “good-enough”) results?” Yet since computers are cheap while robots are correspondingly expensive, low-skilled workers whose abilities include excellent dexterity and eye-hand coordination may be the last to be displaced. Eventually, though, even those jobs will be at risk.
The next decade is also likely to see inroads in the development of AI which will improve AI’s current lack of common sense and understanding. Initially, we will argue about whether it is real sense or understanding, but either way, the result is AI will begin to approach human levels of ability in more areas.
Many businesses are already considering which jobs will be immune — but that’s my point. We won’t be thinking in terms of jobs which might be superseded by AI, but rather which characteristics will make us irreplaceable. We can’t all be artists, dancers or massage therapists. As AI approaches artificial general intelligence (AGI), many more professions are likely to be at risk.
Related Article: Artificial Intelligence Promises to Lighten Our Boring Workloads
What Kind of Future Do We Want to Build?
All of this suggests we need to start building an economy which can more easily adapt to changing technology and the corresponding change in employment needs.
Businesses too must play a part. According to Ellyn Shook, chief leadership and human resources officer at Accenture, "Business leaders must take immediate steps to pivot their workforce to enter an entirely new world where human ingenuity meets intelligent technology to unlock new forms of growth."
To that end, businesses will need to assess the current skill sets and contributions of its workers, as well as their willingness to learn how to use new AI tools. Team leaders must create policies and training protocols that help employees learn new technical skills, while simultaneously seeking their input on how to rethink and restructure workflows. HR departments, for their part, will need to understand how new AI technology works and what soft skills will be needed to enhance the potential of AI.
As AI advances, we can expect a karmic result — as we sow now, so shall we reap. AGIs will be goal-directed systems which learn from ongoing experience. Rather than being programmed, they will be educated and trained like children. Those business leaders who recognize this and begin preparing their workforces now with the right skill sets and AI know-how will be best positioned to reap the benefits of AI and eventually AGI for years to come.
Doing so will also better enable our economy to gradually transition to one which is able to address the ethical dilemmas presented by AGI — such as the displacement of an increasing amount of workers — while taking full advantage of the benefits AGI will bring.