The Problem with Playing the AI Imitation Game

Sophia.

Most people who read that name cannot prevent their mind from turning to the incredibly lifelike AI-enabled humanoid robot from Hanson robotics. And with that celebrity level of fame comes perks, such as the world’s first robot citizenship – in Saudi Arabia of all places. But what’s all the fuss about, really?

Meet Sophia: The first robot declare a citizen in Saudi Arabia

The citizens of tomorrow

Most people have chalked this recent development up to nothing more than a PR stunt – an opportunity to draw attention to new innovation policies in a country trying to turn the page on an oil-focused economy. However, PR stunt or not, it’s kicked off a domino reaction that has just seen Japan become the second country to grant citizenship to a bot, this time a chatbot named Mirai.

This has stoked the fires under an already bubbling pot of contentious conversations over whether we are moving towards a dystopian future – the likes Stephen Hawking and Elon Musk keep warning us of – or not. After all, it raises a lot of questions: can a robot citizen vote? Who is actually voting, the bot or its manufacturer? Do their citizenships extend to their offspring? And since they will be able to “procreate” faster than humans, wouldn’t this pose a risk of overrunning the organic members of a society?

These are all good questions, but everyone has been so concerned with how our treatment of robots will shape them, and whether this will enable them to eventually enslave us iRobot style, that no one is paying attention to how our treatment of them is shaping us.

Everything you can imagine is real

By 2020, 85% of customer interactions will be managed without a human, according to Gartner. However, many people won’t know it. In fact, a recent research paper revealed that both man and machine had trouble differentiating between real Yelp user reviews and those created by AI.

I can’t speak for the bots, but from the human perspective, this inability to differentiate has its roots in early childhood. Humans are ultra-imaginative creatures that begin creating and constructing alternative realities from a young age; every child names their toys and can feel real emotional trauma if anything happens to these inanimate objects. Now, how do you imagine a child would view Sophia? As a robot or a human? Obviously, our AI-enabled future comes with a minefield of potential misunderstandings.

Of course, this isn’t a dilemma that is exclusive to the pre-adolescent. As a celebrity, Sophia has been making a lot of press calls, and though she is a robot without “real” emotions, her human show hosts displayed obvious discomfort treating her as such.

Jimmy Fallon looked uncharacteristically nervous during his encounter with Sophia, as though the robot might laser-eye him if he offended her. Meanwhile, a Good Morning Britain host commented that it was “disconcerting” meeting the robot and that she felt “weird” about her co-host, Piers Morgan, being rude to her.

That AI tech is necessary for the future of every facet of our lives is undeniable — from artificial narrow intelligence to artificial super intelligence. But for now, it does come with its challenges, and this is one of them.

Machine see, machine do

How can you treat something that looks and acts human as anything but that? How can you interact in a world where you can never be entirely certain of who or what you are interacting with? So, while chatbots will soon be an accepted and even expected norm, not knowing if the colleague you are talking to is human or not could be harder to come to terms with, and even more so if this tech imitates more than just the way we look.

AI’s primary function is finding patterns in data and applications for those patterns, but it’s learning how to do this from us, often acquiring and assimilating our pre-conceptions along the way. Just ask Microsoft’s teen-girl chat bot that had to be deleted last year, after interaction with the Twittersphere turned her into a vicious anti-Semite. The lesson? Give a piece of tech an identity and it will develop the flaws to go with it.

Artificial intelligence over artificial imitation

While Sophia is yet to embark on a rant (though as Musk has hinted, that could come in time depending on what input she is fed), her acceptance of citizenship to a country where she as a robot would enjoy more rights than a female human being, already speaks volumes about the dangers of attaching an identity to a piece of technology. And for what? What does citizenship achieve for AI? Not much.

It does generate a lot of publicity for the manufacturer, who – as many people have pointed out – has created a robot that is more impressive for its animatronics than its AI at this point. And this brings me to the greatest downfall of tech designed to imitate us.

I’ve touched on my scepticism about Sophia before. And now I’ll tell you why: She’s not going to transform your business anytime soon, and there will be a mountain of far more worthy data-applications to come before she can. So why then is her name the word on everyone’s lips? Why do you hear more about her in mainstream media than using AI to better measure and optimise marketing in real-time, enhance HR processes, and even generate new molecules for manufacturing? Simple: She looks good on a magazine cover.

The only ends that imitation achieves in tech is to sensationalise it, which then overshadows far greater strides in the industry, capable of actually stimulating positive change. Any machine can be taught to process data. It’s what you do with it that counts.

Do you agree that when it comes to AI, effective, integrated and progressive functioning beats having to engage with a face every time?

About Lee Naik

Lee Naik was named one of LinkedIn’s Top 10 Voices in Technology and is recognized as one of SA’s leading digital and technology transformation experts. Lee is presently the CEO of TransUnion Africa where he leads a portfolio of businesses that help businesses make more informed decisions and consumers manage their personal information to lead to a higher quality of life. 

Check out his Linked blogs and follow him on @naikl for his latest commentary.