Interview with Alexandra Ebert
summary
Claim: no model is unbiased but the target is to identify and remove unwanted or unintentional biases, not chase impossible neutrality. Practical route uses curated non-bias datasets, continuous auditing and mitigation throughout the AI development lifecycle rather than bolt-on fixes, and growing no-code and generative tools to broaden literacy. Industry remains early in adoption, the skills gap is shrinking slowly, and progress is incremental rather than solved by a single tool.
AI-written summary · about
chapters (4)
- 0:00Start
- 0:12What happens if a model is already taught through biased data can we make it more non-biased?
- 0:55How far along in your experiences do you think companies have actually used these non-bias
- 1:48Do you think the same thing can happen with AI now that we have access to all these generative AI tools and in general other AI tools will this technical Gap lesson as we go further?
related talks
20mThere is no such thing as bias-free tech. How models shape our world.
26mBuilding ethical AI: Transparency, bias mitigation, and trust
35mZero Bias AI
56mAn AI with an Agenda: How Our Cognitive Biases Leak Into Machine Learning
05mToward Ethical AI
19mEthical Bias
34m(Un)Ethical Machines? Why AI Bias Is a Problem and What to Do About It
47mEthics & AI: Identifying & preventing bias in predictive models