The Human Touch in Artificial Intelligence
The artificiality of artificial intelligence is overrated.
Brenda loads up an image, and then uses the mouse to trace around just about everything. People, cars, road signs, lane markings — even the sky, specifying whether it’s cloudy or bright. Ingesting millions of these images into an artificial intelligence system means a self-driving car, to use one example, can begin to “recognise” those objects in the real world. The more data, the supposedly smarter the machine.
Since 2016 I have been consigliere to the Turkish startup SOR’UN, who set out to transform the way that humans and corporations communicate with each other. The founders started out building an instant messaging platform that aggregates corporate call centers and provides a smooth-sailing user experience for customers. I loved their idea (I hate waiting on customer service calls), but put simply, I was concerned…
In the years that followed, Turkey went through some tough times. A controversial coup d’etat attempt and its aftermath rattled the business landscape. The currency lost around half of its value, and sustainable recovery remains an optimistic expectation. “Debt” and “bankruptcy” became regular subjects in news and conversation. The chronic “brain drain” turned into a deluge as skilled multilingual knowledge workers simply packed their bags and left.
In this climate, SOR’UN scaled itself from scratch into a pillar of the Turkish tech business ecosystem, generating millions in revenues from long-term partnerships with giants of banking, telecommunications, automotive, and retail, with AI at the forefront of their value proposition.
This, they have accomplished not by going for “one AI to rule them all,” but by creating tools that their own specialists use in crafting bespoke solutions for clients, which themselves allow software and humans to work hand in hand. Even though AI and automation are at the core of their offering, the human touch in their technology is so powerful, I feel like their “chatbots” acquire the personalities of their creators.
Though they set out with different motivations and in vastly different ecosystems, the successes of SOR’UN and Samasource are due to the founders’ common vision on AI, which has nothing to do with being AI-first or how AI is the new electricity. Instead, both companies turned into successful as AI businesses because they nailed this basic insight:
Artificial Intelligence = Aggregate Intelligence
Embedded in every single working AI application today is an outrageous amount of person-hours spent creating examples for the machines to “learn” from. What we call “artificial intelligence” might as well be called “aggregate intelligence,” as it is merely an aggregation of countless human decisions into one decision-making system.
On a strategic level, many businesses looking to “catch the AI train” start by looking to deploy AI for their purposes through some sort of concentrated engineering effort. Whether they are aware of it or not, in this way, they are building on the assumption that it will be somehow better to have machines making certain decisions and creating value for their business, rather than humans.
Instead, SOR’UN and Samasource have chosen to capitalize on a human foundation, adopting a different philosophy: that the value in AI, or any other machinery, comes from the human hand.
Call them Marxists, if you want. But if you ask me, those who underestimate the human touch in creating AI will be paying those who don’t.
I’ll conclude with a quote from the article, and my favorite piece of commentary on it:
Then of course, there’s a question of what happens if the work is no longer needed. Samasource’s main business is, after all, in providing data for automated systems. What if the process of creating that data becomes automated as well?
“That’s the billion dollar technology question that everyone is paranoid about,” Janah said.
“I think there’s a lot of hype around that. But if you actually talk to data scientists, the minds behind these algorithms, you’ll find the machine is much further behind than most people realise.
“We’re going to need training data for a long time.”