It’s been one of the hot topics of 2016, and looks set to continue into 2017: The robots are coming to take our jobs.
All this has been happening against the backdrop of the hype over Artificial Intelligence, which is fast approaching fever pitch. And the world of B2B marketing is certainly not immune. Interest in AI for B2B marketing was already high in early 2016, and skyrocketed with all the hype surrounding Salesforce Einstein and its launch during Dreamforce ’16.
With the simultaneous craze over virtual reality similarly beginning to peak, 2016 was the year some pundits raised the white flag and “accepted”the inevitable: robots will gradually take our jobs, destroying the economic system as we know it.
There’s a lot more to this than pure “hype”, of course; Artificial Intelligence really is revolutionizing B2B marketing, as it is many other industries. And we’re only at the very beginning of the process, so who knows what the future could bring…
At Leadspace, we harness the power of Artificial Intelligence in our Audience Management Platform. Machines are simply capable of processing and making sense of infinitely more information than humans. When you’re faced with a database filled with tens of thousands of leads — let alone millions — there’s just no way any marketing department can effectively comb through and use all of that data to accurately predict who will or won’t buy.
And yet, for all of AI’s clear advantages, the notion of robots “taking over” is pretty much as far-fetched as ever.
Robots can learn, but they’re still missing something…
We’ve touched on this before, but AI, while “intelligent”, isn’t human. And that makes a real qualitative difference.
Even putting aside the fact that “deep learning” still has a long way to go to reach the kind of capacity the doomsayers are so worried about, robots won’t possess human intuition even then. They learn from data — in fact, they’re better at that than we humans are. But some decisions require more than problem solving.
Robots can’t innovate or think up new ideas. They can’t think out of the box, never mind engage in such human professional skills such as negotiating a sale, building professional relationships, and so on. They’re also incapable of departing from linear “logic” when necessary.
Even more fundamentally, behind every technology — no matter how sophisticated and advanced — sits a human to either manage it directly or offer support in the event that it fails or otherwise goes wrong. Leaving machines to themselves without human oversight is like playing Russian roulette.
A black box predictive model runs your data through a machine learning algorithm and gives you a set of results. In the case of predictive analytics, that result would be lead scoring, segmentation, or an ideal customer profile, for example.
But a black box model won’t tell you how or why it came to those results. That’s a major problem — and not just because no astute marketing or sales professional would ever be willing to blindly follow an algorithm.
The fundamental problem with black box models is that your data can in some cases be mistranslated. On a number of occasions, we’ve seen customers looking to break into new verticals who simply couldn’t get the job done with black box models.
Consider the following scenario (one which our own demand gen team successfully overcame not that long ago): your company traditionally sells business services to a particular vertical (“Vertical A”), and now you’re looking to expand into other, new verticals (B and C). If you use a traditional predictive analytics platform to help direct you by scoring Vertical B and C leads, you will almost certainly find that they’re all scored ridiculously low, delivering zero value to you.
Why? Because the machine learning model “correctly” recognized that historically, you’ve done far better with Vertical A companies. Of course, that piece of data is very misleading: you’ve only done better in that space because that was the focus of your business at the time! But it still isn’t “wrong” as far as the algorithm was concerned. That’s what I was referring to earlier — machine-learning algorithms think linearly, while humans are able to think on our feet.
There are other problems with black boxes too. For example, in the long-term you want to be able to learn from your results — not only what worked or not, but why. But if you can’t look under the hood, so to speak, you’ll only at best be able to reap the immediate results without ever knowing — or being able to replicate (or prevent) — the factors behind a particular success (or failure.)
We’ll just create more jobs
The doomsday predictions surrounding Artificial Intelligence all seem to be forgetting one thing: we’ve been here before.
The industrial revolution, printing presses, modern-day farming methods — all of these raised fears of mass unemployment. And in fact, people did lose their jobs. But new ones were created.
As mentioned above, machines need repairs, maintenance, support and direction. But the potential for job creation doesn’t stop there. In the medium- to long-term, AI, much like every major technological advance before it, has the capacity to create jobs we’ve never even heard of yet.
Just think of the various IT and content solutions used to write, post and host this blog. Then consider the social media platforms, search engines and devices you can share and read it on. For that matter, consider the various B2B marketing and sales functions so many of us work in today. None of these concepts even existed 50 years ago, yet all are products of the wave of technological advances which many people (like the Luddites, for example) were convinced would rob workers of their jobs.
We won’t let them take over
OK, this isn’t so much a reason why robots won’t take over, as much as why it doesn’t have to be that way.
While Hollywood’s obsession with robot apocalypses has helped encouraged the assumption that scientists are stupid enough to create robots capable of turning on them, the reality is quite different. No one has an interest in destroying the very economy which makes this kind of innovation possible — and everyone has an interest in maintaining it.
Much as mechanized farming didn’t end up killing off the working classes — on the contrary, it has improved all of our lives significantly — Artificial Intelligence won’t rob us of our livelihoods, but will simply create more opportunities in the long run.
And ironically, the more the potential of robots somehow upending our entire economic system by triggering mass joblessness is discussed, the less likely it is to happen. Awareness of the potential problem — real or exaggerated — encourages greater responsibility.
So perhaps we should just humor the prophets of doom; after all, they just make it even less likely to actually happen.
Picture credit: iStock