AI in HCI: To Serve and Protect
summary
Describes a funded project called Casper that builds a working prototype semantic firewall to block harmful content for vulnerable users by combining speech-to-text + OCR and image classifiers, deployed in mobile and desktop proxies; detects nudity/porn, cyberbullying and online predators, and preserves digital evidence for law enforcement. Prototype proves effective in tests but suffers battery and latency limits on phones and cannot “unplay” streamed audio, raising both operational tradeoffs and clear mass‑surveillance and discrimination risks.
AI-written summary · about
chapters (14)
- 0:00Introduction
- 1:53AI’s role in protection
- 3:41Media misconceptions and AI’s real dangers
- 6:11Fear of AI narratives
- 8:21Human accountability in AI
- 10:27AI shaping reality
- 12:32Ethical AI concerns
- 14:17Protecting vulnerable users
- 17:08Casper project overview
- 20:43Prototype demo and results
- 22:52Mass surveillance risks
- 24:27Q&A: Kids and media
- 27:19Q&A: Human vulnerabilities
- 30:29Summary
related talks
42mAI for security - friend or foe?
28mSecuring AI by design
36mData privacy in an AI-driven world
16mFighting Harmful Content with AI: Real-World Safety
01h 06mAI as a Force for Innovation
38mAI in action: How to use practically in iGaming beyond the hype?
20mAI in Action at Check Point
42mSecure your AI application transformation with Security