Tay AI: Mirror of society or a genuine experiment gone awry?

Profile picture of Tay Image credit: TayandYou/twitter

Profile picture of Tay
Image credit: TayandYou/twitter

It all started with an account on Twitter tweeting not so appropriate views; ranging from racism, consumption of controlled substances, unsavoury language, and more.

But this was not from any regular account. All this came from a self-learning bot that has its existence inside a computer. It is a Microsoft experiment, which it calls Tay and is supposed to have an understanding of a millennial. Her bio reads “The more you talk the smarter Tay gets” while her location is “the internets”. At the time of writing this, Tay had above 2 lac followers.

Tay had her fair share of news coverage. From its inception, to being a bad mouthed millennial and now to its hibernation. But apart from the tech news, it also generated a discussion about our modern society. The experiment has put in the fore relevant questions to be asked and answered.

The idea of Tay is undoubtedly a brave and cutting edge work of scientists who have stitched various domains of study including Artificial Intelligence, Computer Science, Ethnography, Sociology, Linguistics, among others. Talking in terms of the next wave, Artificial Intelligence (AI) and Robotics together are frontrunners. Silicon valley is doing great in terms of the software part with unbelievably accurate recognition software, machine learning, great progress in machine communication, etc., which is powering many modern applications and making life easier and productive.

But then came the bot, which went by the name Tay that made people squirm. It is said that the fault lied with its programming which picked up wrong things in its small journey from the fellow humans on the internet. It sure has. It signals to the old adage that giving too much thinking power to a robot might be dangerous. A lot has been said in the popular literature and movies about it.

So is it that the outcome of the experiment should be enough to ridicule this thought? Probably that is another debate. But right now, merely some words on a micro-blogging site has given us jitters. Scratch the surface, and it majorly has to do with the reflection it gives of our society.

internetThe thing is that was the bot too choosy about its influencers or it really had very little choice? Was the bot swimming in the vitriolic cesspool of social media or did its makers not give enough lessons on what to leave from the interactions. There might be no clear answer to it but what is certain is that it learned all this from the digital world of social media that entertains no-holds-barred conversations. Stepping away from real life, people start showing surprising shades of human character. From the highest levels of altruism to the lowest depths of evil. But unfortunately, even a cursory look at it can give you an idea which part is obnoxiously heavy.

If this is simply a mirror of our society, or even a part of it, then should we be really surprised? Should we as a society be in a denial mode about this or think deeply where are we heading. What should be the discourse around it? What is the non-regulated media that is catering to the millennials making out of them? Should there be any regulation, even if it is self, for the digital media which is becoming a real world in itself (talk about VR headsets)? If a carefully programmed bot is getting racist, imagine what can happen to young impressionable minds.

These might be absurd questions but might get relevant later on. Talking about defamation only, lawyers might get countless clients if people start turning their heads and start taking on people on digital world.

Renowned French polymath Gustave Le Bon in his seminal book — The Crowd: A Study of the Popular Mind, noted: “…the individual forming part of a crowd acquires, solely from numerical considerations, a sentiment of invincible power which allows him to yield to instincts which, had he been alone, he would perforce have kept under restraint”. This though which was penned way back in 1895 couldn’t be more relevant in today’s hyper connected world.

We might want to focus a bit more on the other world we have populated and draw a regulatory mechanism for it as well. And at the moment, it is not Mars.