Robots should not reproduce our mistakes – teaching about equal opportunities and AI

I was an invited lecturer to the introductory course for Technical Mathematicians at the Royal Institute of Technology a few weeks back. The students were really great, and thet had really many interesting view points and reflections related to Equal Opportunities. It was indeed a well worth effort to go to Stockholm and to meet them and to discuss this with them!

We did the course as a flipped classroom experience where the students watched three TED talks before the lecture, and we discussed these during the lecture – and of course other things. One of the TED talks posts was on how to keep human bias out of AI, by Kriti Sharma and the students very much appreciated this.

I think many people have the image that AI is neutral, and that it is disconnected from equal opportunities. Unfortunately this is not true, and AI can indeed be carriers of strong bias. Or as in the below Ted talk where Kriti Shama concludes that the “Poor robots suffer from sexism”, and that the newest technologies are created to reflect a society that is strongly biased.

Robots should not reproduce our mistakes and we need to be careful when launching AI systems in decision making. The problem basically lies in the material that we provide robots with, and how diverse that material is or isn’t. Also there is sexism in what the robots do, and helpful robots that support us are for example often women.

Of course Kriti Shama’s TED talk also provides some ways forwards. Her conclusion is that we should be  

1) Be aware about the biases around us .

2) Give diverse experiences to learn from

3) Make diverse teams build the technologies