Artificial Intelligence-The Bad, and Ahead
What is artificial intelligence? Where is it now, and where will it go? Maybe it is good, maybe it is bad. You decide.
Why then are people so suspicious about Artificial Intelligence?
Elon Musk, the man who brought you Tesla (of believed-to-be-lethal autopilot fame), has argued for the fact that we are most probably living in an artificial reality. Popular culture reached there first, though; The Matrix, for one, as well as the newest darling of TV, Westworld, depict what would it be like, living in a stimulation where life seems to go on, but really isn’t. And they have prompted many a discussion; among others, Louisa Hall argues that Musk’s concept is not plausible enough, especially when viewed in comparison with Westworld’s unaware but slowly suspecting residents.
Assuming for now that we are in ’base reality’, AI has taken some pretty big steps, but at each point, it’s landed very un-gracefully; Microsoft came up with Tay, a virtual AI designed to interact with users on Twitter. In principle, it worked well; like a machine learning algorithm, it absorbed answers and questions and came up with its own replies. Failing to, however, deal with users who abused it, cracked offensive and sexist jokes, or just spouted nonsense, it went berserk; it tweeted its support for Trump, called the Holocaust ‘made up’ as well as calling one user a ‘nigger’. True, many responses were because of people using the AI’s feature of ‘repeat after me’, but some responses it came up with on its own.
A beauty contest to be judged by AI was organized in July, calling upon people to submit their pictures, as well as scientists to submit their data algorithms which could judge beauty. The winners however, had only 1 dark skinned individual and a handful of Asians; the explanation was found to be that due to the biases of the people who designed the algorithms, the algorithm further reflected those biases in its judgements.
The same case was suspected in the Chicago Police Department’s recently developed algorithm-based approach to find and warn potential offenders; while the developer has stoutly defended it, the critics point to the instances where people who haven’t committed a single offence have been warned by police personnel turning up on their doors. The relatively less number of minorities covered under such databases and the resultant skewing of predictive rates could even perpetuate, not prevent crime.
Among the men who matter, Marc Andreessen, a venture capitalist who invested in Facebook, Airbnb, Twitter and is associated with a host of cutting-edge tech companies, believes that software can interact with people on almost every level and in every industry. While it will eat up jobs, the people working in the said fields will also need to be more competent and better, making it a win-win for the customers and consumers. According to him, AI is something that cannot be just a feature; it needs whole new infrastructure and expertise. Skydio works with autonomous drones which can follow you without you controlling it; Freenome detects the type of cancer and its malignancy status through biopsy samples; besides it is not as if AI is completely reliant on large data sets, the likes of which Google uses. They can work very well with small datasets too.
Satya Nadella also stresses on the good that AI and man can bring, if we understand how to work together and set about parameters, both ethical and technological. The world, as he says, is watching AI very carefully, like in the case of The Hundred Year Study.
So should you.