Master of the Game

Until now it has been the stuff of science fiction, movies and nightmares. With the startling advances in artificial intelligence (AI) over the past two-three years, could we soon reach a stage where we would be deftly manipulated by the technology we created? The answer is, almost certainly, yes.

After outsmarting humans on so many fronts and starting to master the domains that humans believed were theirs alone (such as creativity), smart machines are likely to rule in the near future. But it is quite frightening when humans fall prey to scheming robots.

Whether we like it or not, AI is already being used to manipulate people, mostly to get them shopping. Was there an occasion when you were about to buy an item online but left it in the cart as you were having second thoughts? The next thing you knew was a bunch of notifications landing up, saying things like "Come on, you know you want it!" Or "(Your first name), you know you need it!" In fact, one's entire time is spent online on platforms dominated by this sort of manipulation.

Now imagine that a chatbot is used to make the process more engaging and human-like. By conveying bits and pieces of information about you, the chatbot could hold a lively conversation and steer you towards making the purchase.

To take this up a few notches, imagine a robot engaging with a person in an attempt to persuade him/her. It was the subject of a long-ago study in which volunteer subjects were asked to press a button to turn off a computer cat. But the cat begged not to be killed off and persuaded them not to switch it off. A recent study held at the University of Duisburg-Essen in Germany and reported by PLOS One, had 89 volunteers interacting with a robot called Nao. The subjects were 'helping' the robot and seemed to develop an emotional attachment to it in no time. When researchers asked the subjects to turn off the robot, quite a few were unable to do it, especially when the robot used body language to strengthen its verbal pleas. Quite clearly, humans have quickly anthropomorphised the robot - as they have been doing with virtual toys - and are reluctant to harm it, having been emotionally manipulated.

Begging not to be turned off is probably not as disturbing as the possibility of a robot like Sophia, the world's first robot citizen, trying the same trick. Such a scenario was played out on a TV show called The Good Place. There, a pretty humanoid called Janet asked two people to press a switch and destroy her. But every time they approached her, the robot screamed and pleaded to let her go. As they backed off, the robot switched to explaining that she was not human and would feel no pain. It continued until Janet was finally 'murdered' and the manipulated humans had to flee the 'crime' scene.