Machine translation fails typically have comical results, but sometimes the mistakes can have much more serious consequences. Recently, a Palestinian construction worker was mistakenly arrested after posting an image of himself on Facebook next to a bulldozer with the words “good morning” written in Arabic. Israeli police thought he posted the words “hurt them”.
The man was arrested on suspicion of incitement after Facebook’s automatic translation feature changed the meaning of what he had posted. No Arabic-speaking officer read the post and the officers relied entirely on Facebook’s machine translation. There is only a single letter difference between the Arabic phrase “good morning to you all” and “hurt them”. He was quickly released when the mistake was discovered.
Facebook told Mashable, “Unfortunately, our translation systems made an error last week that misinterpreted what this individual posted. Even though our translations are getting better each day, mistakes like these might happen from time to time and we’ve taken steps to address this particular issue. We apologize to him and his family for the mistake and the disruption this caused.”
This mishap comes just a couple of months after Facebook announced that all user translation services will now be powered by neural networks, which are a form of artificial intelligence (AI). Facebook’s convolution neural network (CNN) evaluates groups of words, rather than looking at words one at a time with the goal of translating them quicker and more accurately.
To read more, please see: Facebook’s automatic translate lands Palestinian man in jail.