AI is the latest buzzword on the internet these days. Practically every large tech company is deeply entrenched in the field of artificial intelligence and machine learning. And considering the perils of fake news, and the investigations around this menace involving Facebook, it needed immediate attention.
After releasing its Q1 earnings report, Facebook CEO Mark Zuckerberg took questions, and among them were queries on the role of AI in curbing hate speech and harmful content. Zuckerberg’s reply was interesting and garnered attention. According to a report by Engadget, he said, “It’s much easier to build an AI system to detect a nipple than it is to detect hate speech.”
The response does seem absurd initially, but considering the models upon which artificially intelligent systems rely, such as pattern matching, it always helps to have a visual reference matched with patterns visible in other visual material. In reference to Zuckerberg’s reply, a jot in human anatomy as small as a nipple could be identified given image recognition capabilities of modern systems, however small a ‘needle in the haystack’ as it may be referred to in technology jargon.
WATCH: Huawei P20 Pro Review
Speech on the other hand is subjective to a very great extent, because of the nuances involved. One man’s food is, after all, another man’s poison. A good example being a succulent juicy burger could incite rage in a rather fragile audience. In case of human response, it is of significant challenge for algorithms to detect potential response given the varied preferences in thought within human society.