What Security Measures Are Necessary to Trust in AI?
summary
Panelists describe concrete measures: supervised ML in finance producing a 90% probability customer score that needs business-in-the-loop explanation; anomaly detection for tax data with explainable models and retraining; a research platform with three autonomous cars ingesting about 5 TB/day of sensor data that requires model-based testing, offline-only learning in production cars, and recorded sensor traces for post-incident attribution. Recommendations favor sector-specific rules (automotive safety, financial fraud, personal-data fairness), stronger developer governance, tokenization/PCI/GDPR compliance, and funding for an “Artificial Intelligence Mission Austria.”
AI-written summary · about
chapters (13)
- 0:00AI trust and security panel intro
- 1:13Defining trust in AI
- 3:01Responsible AI and citizen AI
- 7:20AI in automotive safety
- 11:11Government AI regulation strategies
- 16:10AI in tax technology
- 20:05Payment security and fraud prevention
- 24:04Trust versus black box AI
- 31:00AI regulation needs
- 36:20AI explainability approaches
- 40:15AI security in finance and automotive
- 42:49GDPR and AI business models
- 44:16Explainable AI outlook
related talks
45mHardening AI/ML Systems - The Next Frontier of Cybersecurity
01h 05mNavigating the Security and Privacy Landscape of Modern AI
31mSecurity Considerations for Services Using AI Models
31mMaking Trustworthy AI a Business Advantage
44mAdvancing Cybersecurity with AI and Machine Learning: Opportunities and Challenges
35mRisks and Mitigations for a Safe and Responsible AI
38mTrustworthy AI: The Key to Unlocking High-Stakes Applications
24mHow to Build Trustworthy AI