Responsible AI Exam

This fall semester, I’ve been taking the Responsible AI course at University of Agder. A quite interesting semester, studying a topic that is increasingly relevant to us all. For me, it has been an eye-opener in many ways. I knew how these models worked, and the concepts of the technology, but I honestly didn’t give the process more thought - until this fall.

In our group project report, we examined three sandbox projects, where each instance used AI in one way or another. Some focused more on the models themselves, others on the way data was processed - and of course, the effects that would follow. Increased efficiency, more relevant content, better services and experiences.

My deep dive in the group report, ended up being in biasing. Biased data, and how that corrodes into the models. Super interesting, and something I will have in the back of my mind. We humans are naturally biased, but put in a system, it can easily turn into discrimination or loss of opportunities towards others. The systems are argued to remove biasing, and result in more fair outcomes… But when the models train on biased data previously entered by humans, it will inevitably become biased itself.

Fingers crossed for the exam results. I think we did alright, but these group projects can be tough. Will update you once i know more.

Abbi

Hi! I like games, and think that they’re the ultimate form of art; combining music, color and story - all in one product. I work as a Senior Engineer in one of Europe’s biggest energy companies. I also study towards a masters in digitalization.

Next
Next

WoW Forever launched