Securing AI-Driven Finance
summary
An on‑the‑ground account of failures and fixes: a leaked API key lets an attacker probe occupancy patterns; a compromised developer account corrupts transaction pipelines and bypasses fraud detection; attacks include data poisoning, model extraction and black‑box adversarial inputs; cloud risks show up as Kubernetes misconfigurations and vulnerable container images. Concrete defenses are named—zero trust across AI pipeline, model versioning and governance, data lineage, layered infra controls—and examples of correlated detections (network spikes + model drift) for incident response.
AI-written summary · about
chapters (11)
- 0:00Introduction and Speaker Background
- 0:26Securing AI-Driven Finance: An Overview
- 0:38Real-World Example: AI Vulnerabilities
- 1:39The Evolution and Risks of AI in Finance
- 3:26Adversarial Attacks and Data Poisoning
- 5:15Cloud Vulnerabilities and Complexities
- 7:14Human Factors in AI Security
- 8:10Building Resilience: Strategies and Pillars
- 9:03Explainable AI and Multi-Layered Defense
- 10:18Governance and Future Challenges
- 11:33Conclusion and Final Thoughts
related talks
50mBuilding AI Security In: MLSecOps in Practice
18mCommit Secure AI: Embedding Security in Code, Pipelines, and Production
37mFundamentals of AI safety and security
45mHardening AI/ML Systems - The Next Frontier of Cybersecurity
44mImplementing AI in Critical systems
57mAI Applications Security Puzzle
34mInvisible Infiltration of AI Supply Chains: Protective Measures from Adversari...
17mHands-on Practice