Artificial Intelligence is coming to all spheres of life. Can it be biased?
Artificial Intelligence is a popular new trend. Scientists are working hard find as many ways to use it as they can. The logic of computer brain seems to empower anything it is applied to but is it really so?
Let’s start with simple old stories. Scientists have recently discovered that there were strong associations, known as word “embeddings”, between European or American names and pleasant terms, and African-American names and unpleasant terms.
In a paper about the new study in the journal Science, the researchers wrote, “Our findings suggest that if we build an intelligent system that learns enough about the properties of language to be able to understand and produce it, in the process it will also acquire historical cultural associations, some of which can be objectionable.”
Try searching Google for “Beautiful people.” How many Black faces will you see? To make sure bias exists you can try searching “three black teenagers” and “three white teenagers” respectively, as Kabir Alli did. The first search showed disturbing images of black “thugs” and “convicts”. But when he then replaced the word “black” with “white” in the search field, happy smiling faces of young men and women popped up.
Now let’s talk about something more serious…
Beauty.AI 1.0 in the January of 2016 launched an international beauty contest with over 6,000 participants from over 100 countries. The competition was judged entirely by artificial intelligence. There was only one winner with dark skin among the winners. Can you imagine that the same Artificial Intelligence mechanism selecting U.S. Senators, Nobel laureates, or U.S. presidents? What if the same principle will command some futuristic Robocop to shoot (or not to shoot)?
The problem is that racist people can’t create non-racist technologies. Just remember Tay.