Opinions

View: Will I pass Elon Musk's 5% feku test?


While the nation has been shivling in the heat, Elon Musk has forced me to ask myself the deepest of deep questions that lurk in the deep end of the pool: What am I?

No, not that sort of inquiry where to find an answer, I have to engage with a suave happyguru, who replaces Management 101 words with key Sanskrit ones to come up with gems like ‘You hold the astika to your inner nastika’. My question is: Am I human?

Ever since the Musked crusader cast doubts on the number of real humans out there on Twitter, I have started to wonder whether I am a human, or a bot masquerading as one. Or worse, like the replicants in Bladerunner – or Bengalis who think they’re Englishmen just because they talk in Wren & Martin — a bot who has no clue that it’s a bot and believes itself to be a human?

The signs have been piling up for a while now. Whenever I ask Amazon’s slavechick Alexa to play a particular song, say, Abba’s ‘Lay All Your Love on Me,’ it plays me something else, say, the remixed version of ‘Laila Mein Laila’ from Raees. Then, while I was scrolling to order an old copy of Desmond Morris’ Manwatching: A Field Guide to Human Behaviour, the algorithm threw up a ‘suggested purchase’ for a 4-pack Manforce Cocktail Condoms (Strawberry+Vanilla & Chocolate+Hazelnut).

It’s only after Musk announced that he wants proof that not more than 5% of Twitter users are spam accounts, before he proceeds with the deal to acquire the company, that the paisa dropped. Machines have been behaving oddly with me – my toaster refuses to golden-brown my bread; dialogues in movies on OTT platforms have lags longer than Deepika Padukone’s legs; my car steering wheel positively left-leans even in today’s right-of-centre traffic…

All because I suspect that these machines have detected that I am no human at all. And like some managers in uppity restaurants who give good service only if you’re a gora, these machines also treat me like I’m one of them (read: as shit). Or worse, a comprador class bot which has picked up airs of its masters.

So I’m trying to find out what I am – one of ‘Musk’s 5%’, or a flesh and blood (read: monetisable) human.

In 1950, the mathematician Alan Turing published a paper, ‘Computing Machinery and Intelligence’ (bit.ly/38AC3Ng), which opened with its now famous line: ‘I propose to consider the question, ‘Can machines think?” Turing used a variation of the ‘Imitation Game’, where three people – a man, a woman and an interrogator of either gender – participate, with the interrogator in a different room from the other two.

The object of the game is for the questioner to determine which of the two is the man and which the woman by asking a set of ‘gender-loaded’ questions such as the length of one’s hair or brand of one’s soap. The object of the two interrogated is to confuse the questioner. (Turing was gay and was hounded for it in post-war homophobic Britain, giving his inclusion of the Imitation Game’s gender hide-and-seek function an ironic, tragic twist).

Now, Turing set the Imitation Game with two humans and one computer, with one of the humans trying to determine which of the two entities inside is a human and which a machine through a set of questions. The Turing Test was created in a time of simpler computers. Today, we have much more sophisticated methods of trying to separate bots from persons.

One such test, the Reverse Turing Test, in which a human tries to convince a computer that he or she or it is not a computer, is the common CAPTCHA, or Completely Automated Public Turing Test to Tell Computers and Humans Apart. I keep muffing up CAPTCHA each time, whether the website asks me to click on image boxes with fire hydrants or trucks or traffic signals – one more source of existential concern.

So, if Musk asks, ‘Is Indrajit Hazra part of the 5% riffraff Twitter fekus?’ frankly, I don’t have an answer yet. The way I can’t confirm whether this column was algorithm-written, or generated by a human with dodgy heuristic capabilities.



READ SOURCE

Leave a Reply

This website uses cookies. By continuing to use this site, you accept our use of cookies.