Cloud Enabled Biometric Data Collection
summary
Claim: cloud-hosted biometric datasets are routinely stored, synthesized and sometimes sold, and that cloud = someone else's computer. Evidence: commercial facial classifiers show up to 34.7% error rate on darker‑skinned females and repeatedly misidentify people, producing wrongful arrests. Tools and fixes mentioned: commercial classifiers, iris and gait scans, plus policy remedies like GDPR and CCPA and ethics frameworks. Built examples include demo knowledge‑graph training (Neo4j Graph Academy) and simple chatbots in Python/JavaScript.
AI-written summary · about
chapters (12)
- 0:00Welcome
- 1:27Speaker background
- 3:15Non-traditional tech paths
- 5:48AI everyday uses
- 7:45Understanding biometrics
- 10:35Risks of cloud biometrics
- 12:02Facial recognition tech
- 13:47Facial recognition bias
- 18:25False match consequences
- 21:14Cloud practitioner actions
- 22:19Regulations and ethics
- 23:41Final takeaways and resources
related talks
27mThe Limitations and Danger of Facial Recognition
37mBiometrics & Privacy: Time to Faceoff or is that FaceApp?
28mAlgorithmic Bias: Data as a Weapon with Ayodele Odubela
01h 00mNavigating the Biometric Anatomy
01h 00mBiometric Unsecurity
21mMaking authentication your competitive advantage under GDPR and PSD2
50mPitfall or Opportunity: GenAI Legal Case Studies Revealing Practical Advice
39mDefensible Application Security for the Artificial Intelligence Era