Join Catalina Butnaru, ambassador for Women in AI at Supercharged 2018 at the panel 'How pc is your AI? Why brand owners should be concerned about bias'
I’m contributing towards more thoughtful, ethical, and inclusive applications of AI, by getting involved in the Ethics in Action working groups at IEEE.
What do you think the impact of AI is outside of the lab?
Add a cognitive layer to any predictive and pattern-recognition software and you will measure immediate improvements in speed, and accuracy. However, even if narrow AI applications are powerful efficiency engines powering any predictive and analytics function, high error-rates could significantly slow down live deployment of AI-based systems.
What’s the biggest impact bias in AI has had to every-day life?
Some forms of bias are hard to detect, as they are usually related to embedded biases in data. We still do not know what the effect looks like in everyday life, apart from those very obvious blunders and uncorrected mistakes that had already received media coverage. Researchers such as Silvia Chiappa are working on ways of correcting biases in algorithms, but still warn us that unless we clearly know what biases decision making looks like, we cannot correct it. At a larger scale, uncorrected introduction of social biases can amplify systemic or artificial unfairness. For example, people with a specific ethnic and financial background will continue to access more resources, jobs and opportunities, despite evidence that people who do not fit the parameters perform equally well or even better.
Eliminating bias in AI: Will we ever get there?
Yes, and first we must define subtle biases, then commit to correcting them.
Which sci-fi film do you most want to become reality and why?
Circle! It’s a movie about radical transparency, which is an important ethical value more companies should pledge to adopt within their practice.
Describe in three words your session at Supercharged!
Fair AI (in) Marketing