Edmond Mellina: I first came across the amazing power of artificial intelligence (AI) in the late ’80s, when I learned to code in Prolog. It’s the AI programming language that now powers IBM Watson’s ability to process natural language.
Those were the very early years of AI. I was on the last stretch of my engineering degree in Marseille, France. Alain Colmerauer, the inventor of Prolog, was a professor at a nearby university. The two local institutions had a partnership. So, we had the opportunity to play with the fascinating language he had developed. I remember being in total awe of the power that a few lines of code could unleash.
Today, we are clearly on the cusp of an AI revolution. The convergence of big data, machine learning and massive computing power is making AI more and more ubiquitous. Our speakers provided numerous examples — from HBO’s Silicon Valley’s silly “No Hotdogs” app to IBM’s Watson Talent.
But for me, the most thought-provoking aspects of the discussion were the couple of either-or scenarios Cindy Gordon painted for us. In each story, we could see a future where advancements in AI could lead to either “a perfect world or a perfect storm.”
As I discovered through my first lines of code in Prolog, AI is indeed hugely powerful. It can either augment human intelligence or create nightmare scenarios. At this stage, humans are still controlling the AI revolution. I hope we still have a decade or so to make the right choices.
Paul Pittman: AI used to be science fiction and recent years have witnessed its evolution to science fact, and now it hovers at the edge of realizable. However, notwithstanding the exciting future laid out before us by Gordon, I fear there is a significant difference between what is possible and what will become practical in the near term.
We are often reminded that smartphones have been around for less than 10 years as evidence that technological advancement occurs quickly. But we forget that dumb phones were around for 10 years before that and stupid phones 100 years before that. With all of their micro technology and “smartness,” they are still only doing things we could have done ourselves, given time, some additional equipment and a library.
Boomers are fascinated by technology; millennials use it as a tool but are equally less willing to allow it to make choices for them. Until we are comfortable with machines making choices for us and decisions about those choices, AI will never reach its full potential.
Technology that we rely on today essentially accelerates tedious repetitive tasks or performs digitally activities that we carried out manually or would have had to had we had the inclination; for example, the daily summary of bodily vital signs maintained by my Fitbit. It’s mostly unnecessary data that uses valuable resources to collect it.
I can’t think of many applications that we allow technology to propose other than mechanical “suggestions” for us. For example, I need to get to Waterloo by 4 p.m. and the fastest route would be X. If I so choose, I can elect to override this and travel via route Y or even walk instead of travelling by taxi. Human nature is fickle. Binary decision-making, no matter how flexible, minimizes humanness and, in turn, minimizes the risk of taking the entrepreneurial out of business.
Jan van der Hoop: I’ve been married 30 years and it’s hard to admit that Amazon knows more about my wife’s preferences (at least as it relates to online shopping — for now) than I do.
Is it a “sad” truth though? Or does it make her life better by saving her time or suggesting complementary items she hadn’t considered? I guess it’s all in the eye of the beholder.
And, as Paul suggests, as elaborate as those algorithms are, like mapping software, it’s all an extension of the old “If X, then Y.” If you’re here and you want to get there, and we know traffic’s backed up along this route, then suggest this one. Or, if she’s bought this and searched using this, this and this term, then suggest that. Elaborate, powerful and clever? Yes. Life-improving but hardly life-changing. And, as we learned, this is AI in its infancy.
For me, Gordon highlighted four huge and largely unspoken implications to organizations:
Ethics: Boundaries need to be baked into the design from the very beginning, not after the fact: the lines that will not be crossed, and the “lesser evil” compromises that are programmed in to the software. My guess is that in many instances, this is given short shrift in the rush to get a product to market.
Judicious management and good stewardship: Using AI in an uninformed way and without a clear plan is tantamount to putting a loaded revolver in the hands of a two-year-old. Organizations must be prepared, informed and intentional in the use of AI.
Data integrity and security are critical: The reality is that most organizations do not have very clean databases. The risk is obvious.
Focus outside, not inside: We tend to be so obsessed about the internal workings of our organizations, but the world is moving around us. Opportunities for growth occur outside the organization. Good AI should help identify those opportunities that we quite simply don’t have the eyes to see.
The big question underpinning the entire debate is how will we take the best of both our essential humanity and the power of AI, and blend them for the greater good?
As perverse as it sounds, our humanity may serve to insulate us from many of the risks of bad technology. We are, after all, fundamentally irrational creatures and I think it’ll be a long time before artificial intelligence can predict our irrationality.
But we place ourselves at serious risk when our intellectual laziness allows us to delegate our responsibility for perception and decision-making. One example relates to how successfully agents of a foreign state, seeking to disrupt and divide, influenced the outcome of a recent election. They did so simply by spreading misinformation that created an emotional reaction that inflamed pre-existing biases and triggers in a segment of voters.
And that was just a bunch of Russian hackers running a misinformation machine. Child’s play. They bought ads on Facebook, for crying out loud. Imagine a world where AI has the power to shape our most fundamental perceptions and beliefs, where we respond reflexively to what we see and hear, but the things we see and hear are a false construct. I’m not sure we are far away from that world.
The debate over AI should be a wake-up call. Governments and organizations must be active and on guard. I believe the skeptical and tough-minded will survive.
Tracey White: I agree, Jan. We are already living in a world in which algorithms are making decisions that impact our lives in real ways. As Gordon warned, we need to take note now because competitive pressure will drive automation. We won’t have the luxury of time to adjust and consider the ethical impacts of new technologies.
Futurist Stefan Hyttfors has said that as fast as we think technological change is now, it will never again be this slow.
What does all this mean for HR? As a keeper of significant organizational data, HR is a nexus. Early technology automated the paper-based activities of existing corporate functions, but it reinforced organizational silos.
The current generation of technology is breaking down silos, allowing bodies of data to be integrated so business decisions can be made from intersecting information flows.
The next generation of technology, as we saw from the Apple and IBM speakers, will allow for computer-driven decision-making.
Artificial intelligence and machine learning will eliminate a lot of administrative activity and, in so doing, they will generate enormous cost savings. The potential for corporate bottom lines will be irresistible. This will mean significant structural change within organizations, including for HR.
HR has struggled for over a decade to operationalize the David Ulrich model that aimed to make HR into a true business partner. He was ahead of his time because the information flows needed to achieve its vision weren’t available. That will change with artificial intelligence.
HR business partners will have access to instant information flows more common to sales and finance. Immediate information will make for more reliable decision-making. It will increase HR’s value in the C-suite, but it will mean the human resources skill set will need to change significantly from a focus on program design and implementation to data analysis and integrated business decision-making.
As stocks of information become flows, and decisions are automated, the HR function of today will not exist in 10 years.
Mellina: That’s right. And the rest of the organization won’t look the same either. So, HR has a dual challenge to tackle: Reinvent itself while guiding the organization through a massive and fundamental revolution.
The arrival of the robots is raising the need for forward-thinking HR leaders. It’s time to step up, we need you.
© Copyright Canadian HR Reporter, Thomson Reuters Canada Limited. All rights reserved.