Issue #59: Making AI less socially biased

A new AI training technique announced by OSU and Adobe researchers

Welcome to Issue #59 of One Minute AI, your daily AI news companion. This issue discusses recent research from Oregon State University.

---

Introducing FairDeDup

Researchers from Oregon State University and Adobe have developed a new AI training technique called FairDeDup to reduce social biases in AI systems. This method focuses on fair deduplication, which removes redundant data while considering diversity factors such as race, gender, and occupation. By thinning datasets in a content-aware manner, FairDeDup aims to mitigate biases that perpetuate stereotypes, such as associating specific professions with certain demographics.

The approach allows AI systems to be trained more cost-effectively and accurately while promoting fairness defined by user settings rather than relying on biased internet datasets.

---

Want to help?

If you liked this issue, help spread the word and share One Minute AI with your peers and community.

You can also share feedback with us, as well as news from the AI world that you’d like to see featured by joining our chat on Substack.