Does your product need non-human intelligence?

July 3, 2017 Typeform

Our vision at Typeform is to “make things a little more human.” replace unnatural moments of interaction between people and machines with the most innate form of communication: the conversation. By leveraging its beauty and simplicity, we allow real people to learn about each other — creating a context where conversational data collection can thrive.

As the man in charge of analytics I wonder: what does this mean for our team? How can we data scientists make things a little more human?

The team’s long-term vision for AI is ambitious: reproduce what’s going on in an actual human conversation. In other words, the constant stream of mutual feedback and the huge amount of information that’s carried by our tone of voice, pauses, looks, gestures, proxemics, choice of words. Not an easy task.

But let’s be honest: today’s AI is more artificial than intelligent.

Modern machines outperform humans in many ways. They’re better drivers, wiser stock brokers and, only recently, stronger GO players. But performance isn’t everything. There’s a much bigger gap to bridge between AI and the human brain, and that’s cognition: “the mental faculty of knowing, which includes perceiving, recognizing, conceiving, judging, reasoning, and imagining”

In all fairness, today’s AI does very little of that. It’s only capable of capturing and mimicking highly-complex cause/effect relationships between several inputs and one (or more) outputs. That’s not intelligence I’m afraid, it’s math. Correlations to be more precise.

For example, neural networks do not perceive, judge, or recognize a face. Nor can they truly imagine one. But they do a great job at spotting distinctive sequences in their numerical representations, and associating them with the most similar ones in a given set of examples, to ultimately return the best match. And they do this quickly and with a 99.9% accuracy. So we’ve got to give it to them.

But there’s more. Human motivations are bizarre and irrational. We don’t pursue optimization, we pursue happiness. And whatever makes us happy is the product of some arbitrary personal criteria — an expression of our personalities that little has to do with achieving the best possible outcome.

Lastly, to make things even more confusing, the human brain is seriously flawed. Our days are plagued with uncertainty, poor judgement, and petty failures. The keys that are not where you just left them, the carton of milk you can’t find in the fridge, that sweet memory that never happened, that thing you forgot which enraged your partner and ended up costing you weeks of chastity…

So here we are: highly intelligent post-apes applying the advanced cognitive abilities of our imperfect brains to make sense of the world around us, so that we can pursue some arbitrary goals.

That’s what makes us human. Can correlations reproduce all that? I seriously doubt it. So what does this mean for companies like us?

Perhaps we should wonder what kind of intelligence our products need. They’re built for people after all, not for machines. We might be on the verge of an AI revolution but that’s hardly going to be human. Is it time to worry yet?

Join 25,000+ people who read the weekly 🤖Machine Learnings🤖newsletter to understand how AI will impact the way they work.

Alessandro Pregnolato is the director of analytics at Typeform. He’s responsible for data operations, data science, and machine learning: empowering Typeform to make the best decisions while promoting a data culture through passionate and effective communication. When not obsessing over data, he’s the lead vocalist and bassist in a rock band.


Does your product need non-human intelligence? was originally published in Machine Learnings on Medium, where people are continuing the conversation by highlighting and responding to this story.

Previous Article
Mini AI crash and smarter drug research
Mini AI crash and smarter drug research

SourceAwesome, not awesome.#Awesome“Twice in the space of six weeks, the world has suffered major attacks o...

Next Article
Winning Strategies for Applied AI Companies
Winning Strategies for Applied AI Companies

Key Success Factors after reviewing over 70 companies that have raised at least $7MAs primarily SaaS invest...