In Russia, AI may mean “argue intensively.”
I was recently on a panel on “The effect that Artificial Intelligence (AI) will have on jobs”. The conference was in Moscow, and the panel included AI experts, the HR Director of the Russian Pension Fund, the odd minister and Directors General of various government departments, the head of a university… and me. Heady stuff.
But then it went downhill. It was a three-hour session, and wide-ranging discussions and heated arguments (Russians seem to like arguing). Problem was we were talking about the wrong kind of AI – the kind that is still far in the future and may have no impact on jobs at all.
After a while, I leaned across and said to my neighboring panelist, “Am I in the right session? We’re not discussing jobs at all, and we’re not even talking about the right AI.”
He leaned back, “Yes you are, this is the AI impact on jobs session.”
I replied, “Why aren’t we keeping to the topic?”
He smiled, “This is Russia!”
(I like Russians – they are friendly, passionate, intelligent and accommodating – and argue a lot.)
So apart from the entertainment value of watching Russians argue among themselves, the session was not valuable. In my naivete, I resolved to try to do something about it. When my turn came to summarize, (I had made an opening statement, then spent the rest of the time swivel-headed, watching the protagonists at play) I said something like this:
“For the purposes of this discussion, there are four types of AI:
Algorithm and data-based AI – this is usually called machine learning, but let’s stick with algorithmic AI for now.Then there is, let’s call it brain-modeling AI, that tries to think like a human. That’s what we’ve been discussing here, but it doesn’t affect jobs, and it’s a long way off being a mature technology.There is also symbolic AI, which tries to recognize (analog) things – face-recognition is a crude example of this, but usually, it takes a supercomputer to understand the panoply of visual cues that humans know at the age of three.Then there’s swarm AI that looks at collective behaviors of large populations in decentralized and self-organizing systems.
Only the first type threatens jobs today – very seriously. So let’s talk about that. Algorithmic AI follows rules, is deeply repetitive, uses data, and handles the mundane. (The interpreters, bless them, translated all this into Russian without blinking.)
So these are the jobs that are threatened – mundane, repetitive, tightly bound, and rules-based. This form of AI is relatively cheap, and with RPA (robotic process automation) is a compelling proposition for businesses and factories. AI (and RPA) doesn’t earn a salary, complain, get sick, take leave, join unions, need pension funds, and works 24/7 – what’s not to like? Hundreds of millions of jobs are at risk.
But what can people do that AI can’t? Lot’s.
People have curiosity, divergent thinking (about what’s possible), are intuitive and creative. They use judgment and, we hope, ethics and wisdom. They get emotional, have empathy and understand non-verbal cues. They have a cultural background that brings diverse views to the party. They facilitate, motivate, and counsel. Probably most important, people can lead, create a weltanschauung, communicate and, even more importantly, can they take responsibility.
Jobs that use these skills will be relatively safe. However, people being replaced have hard skills, not the soft skills above. Massive retraining will be needed. Let’s talk about new jobs, skills, and training.”
There was silence in the room. Then a collective sigh. We were back on track again.