Bias in Biometrics: How to Authenticate Quickly and Fairly

Presented by

Alix Melchy, VP of AI at Jumio

About this talk

Bias can creep into algorithms in several ways. AI systems learn to make decisions based on training data, which can include biased human decisions or reflect historical or social inequities, even if sensitive variables such as gender, race or sexual orientation are removed.  Facial authentication relies on the unique biological characteristics of an individual to verify that they are who they claim to be, so bias is a huge issue when attempting to verify user identities. At Jumio, we use AI to create deep learning models in order to improve identity verification accuracy, identify fraudulent IDs and make the verification process faster. Join us in this webinar to hear from our VP of AI, Alix Melchy, as he dives into these challenges and presents solutions to help eliminate bias in algorithms.
Related topics:

More from this channel

Upcoming talks (5)
On-demand talks (84)
Subscribers (21854)
Jumio helps organizations to know and trust their customers online. From account opening to ongoing monitoring, the Jumio Platform provides advanced risk signals, identity proofing and compliance solutions that help you accurately establish, maintain and reassert trust. Leveraging advanced technology including automation, biometrics, AI/machine learning, liveness detection and no-code orchestration with hundreds of data sources, Jumio helps organizations fight fraud and financial crime, onboard good customers faster and meet regulatory compliance including KYC and AML. Jumio has processed more than 1 billion transactions spanning over 200 countries and territories from real-time web and mobile transactions. Based in Sunnyvale, Jumio operates globally with offices in North America, Latin America, Europe, Asia Pacific and the Middle East and has been the recipient of numerous awards for innovation. Jumio is backed by Centana Growth Partners, Great Hill Partners and Millennium Technology Value Partners. For more information, please visit