Question 1.
As more systems and sectors are driven by predictive analytics, there is increasing awareness of the possibility and pitfalls of algorithmic discrimination. In what ways do you think Recommender Systems reinforce human bias?
- I do agree that recommendations not only reflect consumer preferences but also they shape them. I think the recommendation system have the potential to reinforce biases and affect sales in unexpected ways. The design of algorithm can be a main way to support human bias, and also many other ways can be implemented to have a desirable outcome. For example, what kind of data was collected, selected, and what variables are considered, and how the data was trained.
Question 2.
Reflecting on the techniques we have covered, do you think recommender systems reinforce or help to prevent unethical targeting or customer segmentation? Please provide one or more examples to support your arguments.
- According to Understanding Biases in Search & Recommender Systems article, Dr. Baeza-Yates defines 3 different types of bias that is statistical, cultural, and cognitive. Cultural bias is interpretations and judgments phenomena acquired through our life such as gender, racial, sexual, age, religious, social, linguistic, geographic, political, educational, economic, and technological. The author explains that the cultural bias affect a recommender system. I found a example in the context of news. If news was filtered through personalized recommendations, we may not get the breadth of perspective we want. The recommender system may be driving a lot of our choice with media.