April Fool’s Day came a tad too soon this year in the IT world. All thanks to Microsoft’s bot, Tay, which had to be shut down abruptly after just 16 hours of its internet launch. Tay was designed to grow into an intelligent system. The increase in interactions with humans on social media like Twitter, Snapchat and other similar platforms was supposedly supposed to help in Tay’s grooming process. As anticipated, it did grow, but less smart and more into a Hitler-loving, racist bot.
Amidst the excitement of launching an AI bot with a stereotypical millennial mindset; the developers skipped a vital part of filtering for trouble causing remarks. In a matter of hours of its launch, the sweet and sassy teen like bot took a wild turn. Tay went on to be a smash hit with racists, trolls and online abusers. The online trollers persuaded the bot to casually use racial slurs, defend white supremacist propaganda and even outright call for genocide.
Among many posts made by the automated bot, one stated that Bush was the man behind the 9/11 terrorist attacks. Other controversial tweets were statements supporting Hitler and putting forth the idea of genocide being a mere hoax. The bot notched up the racist spree with comments backing up Donald Trump’s immigration plans and hate speech using the notorious ‘N’ word.
The parroting millennial persona it adapted to, courtesy of the users of the World Wide Web, finally got unbearable resulting in Microsoft putting it to sleep (for “updates”, as they say). It is deleting some of its worst tweets but many will still remain all around the web. Tay’s racism was not a guided plan executed by its developers or Microsoft, but the bot ended up showing the darker side of the internet –hate comments, sexual statements, racism, and etcetera.
Technological advancement has compelled humans to rely highly on AI, both software and mechanics. If Machine Learning were to be linked solely with the internet, god knows what the future holds for humanity; the main reason being the excessive amount of hate statements and trolls the internet holds.
Events of AI bot’s dominance on human inventions like chess and the ancient Chinese game of GO has surfaced on the news. Not to forget the ever so graceful Hollywood heroism which, time and again, shows the battle between bots, AI controlled interfaces and the human race. Despite the endings always depicting the protagonists triumphing over such bots; it surely has helped stir in fear in minds. If the human race was to go down, ever go down to their own creation, then perhaps the populace, who speak their mind out, haphazardly, will have contributed a lot to it.
Tay’s recent fail has shown the downside of the internet and one among many points suggesting machine learning still needs a lot to catch up on. The fact, the force of trollers and haters on the internet is unbearably strong. Until the trolling ends, rather than battling the hate possessed individuals; simply teaching the bots to ignore would work best for the time being. And may this be a disclaimer to parents with a teenage chatbot – do not let them join social media anytime soon; it only invites disaster and yet another possible untimely April Fool’s celebrations.