Microsoft Tay
The idea behind AI chatbots pretending to comprehend manners and sentiments is about as fresh as a concept from yet another movie of Nicholas Sparks. Tech behemoths are madly engaging themselves in a fierce race to create digital assistants and chatbots powered by artificial intelligence.
The speed of creation for these artificial ‘chat-mates’ is increasing and certainly it isn’t planning to slow down anytime soon. From Microsoft Tay, which can turn you into a Hitler disciple, to Facebook’s AI chatbots allowing you to order flowers, and of course, then there’s ever-chatty Microsoft Cortana helping you book a hotel room or call Uber for a ride. It seems as if all the major tech enterprises want to offer you witty, bold, and humorous virtual, AI-powered concierges. Searching or shopping, however, with these AI chatbots is less competent to the usage of standard web tools for the same purpose. Even worse, it absolutely flips the concept of the relationship between the human and a machine on its head.
There are always helpful AI chatbots like Facebook’s recently introduced artificial intelligence-powered bots, Apple Siri, and Microsoft Cortana, but let’s not forget controversy’s favorite child- Microsoft Tay, which lies on the opposite end of the aforementioned names. Microsoft Tay is the best example to demonstrate the ways in which artificial intelligence-powered chatbots go wrong. The abbreviations loving, cheerful, and prim Tay suddenly turned into a sexist, Hitler-loving, perverted, and racist maniac overnight. One of the highly championing AI programs of Microsoft suddenly turned disaster by bashing the company’s image in the worst manner possible on the giant social media network. Merely a day after its ‘talk of the town’ launch, Tay, Microsoft’s AI chatbot was laid to rest by its creators after it brought a massive public outrage by its warp and perverted humor.
Let’s just agree to a point that no matter how intelligent an AI-powered chatbot may be, they’ll always be a waste of time.
These chatbots powered by AI are mere tools, not our friends. We do not need them to exchange bogus pleasantries with us. Because let’s just say, that’s not what we want, seriously. Imagine coming home from a tiresome board meeting where your acquaintances haven’t spared a single chance to piss you off. A certain AI chatbot starts questioning you on how well was your day, or did you have fun or worst, they’re glad you enjoyed your day! This is the last thing you want to be subjected to after a terrible day. In such situations, all you need is to shut them up and continue your work. There’s always some terrible and disgusting aura surrounding the concept of fake pleasantries.
Of course, tech enterprises aim to anthropomorphize us with their artificial intelligence chatbots. IBM has ads featuring Watson, and Apple used to run awkward commercials that featured Zooey Deschanel and Samuel L. Jackson chatting with digital assistant Siri.
Voice assistant of Google is a darling. It’s a noteworthy exception in this race because it never feigns humanity. If you’re interacting with a computer device as if it were a person, at some level, you’re considering it as either a friend or a servant (the latter is considered the most).
No doubt, with drastic advancements in AI, problems such as vocabulary is going to get ironed out over the period of a short span. However, even after this, it will remain easier and faster to interact with a computer using unambiguous and short commands, instead of lengthy voice commands. Pressing the ESC key is always going to remain as a better option than speaking ‘Hey, Chatbot, leave it. I’ve changed my mind.’ No wonder why hardcore developers still prefer working using command line options rather than voice commands. Touching and typing to organize your menu, move files, or write programs will always have a better preference than commanding for each process.
But, if developers are so keen on entertaining us with AI chatbots then they must find some effective options. For instance, interacting via voice can be faster and easier when there are straight commands rather than fake pleasantries.
Instructing your bot to increase the temperature of the air conditioner by 1 degree is a faster and easier way to control than using the keyboard and saying, it feels so hot here, please do something about it. This certainly won’t work.
The personal computers were launched more than thirty-five years ago, and it has been 20 years since the internet went mainstream. By this time, people certainly know how computers work. There’s no reason at all why we can’t interact with the machines in just the way they are rather than wasting time with fake pleasantries via chatbots.
