Fostering Trust in AI-Driven Finance
summary
Argues that trust must be built across the entire model lifecycle, from business case and stakeholder mapping to data, feature engineering, model discovery, deployment and monitoring, with concrete practices: classify features as semantic / derived / non‑semantic transformations, validate features with branch users, and refuse models that users cannot explain — a bank project abandons a fancy neural net and threw away three weeks of work for a simpler model the business accepts. Recommends SHAP/LIME, causal discovery, three‑level monitoring (technical, model, business) and asking users before deploying interaction terms; cites a mortgage pricing tool for mortgage valuation example.
AI-written summary · about
chapters (51)
- 0:00Introduction
- 1:08Trust in AI-driven finance
- 1:52AI development and resistance
- 4:23Trust in AI models
- 6:06Business case to deployment
- 8:11Data for trustworthy AI
- 11:23Model explainability
- 13:13Model building best practices
- 17:32Fairness analysis
- 22:09Model monitoring
- 25:00Future of AI trust
- 25:44Q&A
- 34:26Next speaker Nathan Coyle
- 35:32Peace tech overview
- 40:11Communication in peace tech
- 42:02VR empathy and storytelling
- 44:13Digital literacy in peace tech
- 46:23Cybersecurity in peace tech
- 47:07Early warning systems and bias
- 50:16Government challenges with open data
- 55:37Open data reflections
- 58:05Human-centric open data
- 1:00:10Policy gaps and solutions
- 1:01:32Q&A and closing peace tech
- 1:01:52Responsible AI keynote
- 1:02:35Responsible AI risks
- 1:04:02Consequences of irresponsible AI
- 1:05:28Responsible AI framework
- 1:07:12AI across data science lifecycle
- 1:09:01AI governance and oversight
- 1:11:53Interdisciplinary AI governance
- 1:14:01AI for children compliance
- 1:16:32AI governance challenges
- 1:18:17Red teaming vulnerabilities
- 1:20:44LLM evaluators
- 1:22:51Adaptive AI governance
- 1:25:44Q&A responsible AI
- 1:26:48AI trustworthiness keynote
- 1:28:13AI adoption and regulations
- 1:29:17EU AI Act standards
- 1:31:03Safety-critical AI challenges
- 1:34:15Global AI regulations
- 1:36:01New legislative framework
- 1:38:11ISO standards compliance
- 1:40:21EU AI Act compliance
- 1:42:51AI robustness requirements
- 1:45:03AI system validation
- 1:47:53Explainability and bias
- 1:50:44AI robustness results
- 1:53:57AI conformity tools
- 1:57:07Q&A AI Act compliance
related talks
24mToward Trustworthy AI: Explainability and Robustness
44mPanel Discussion: The BFSI Playbook - Scaling AI with Trust and Quality
45mWhat Security Measures Are Necessary to Trust in AI?
16mTrust by Design: Building Reliable AI for Sensitive Environments
13mMLOps in FinTech: Architecting Responsible AI
31mAI risk governance
45mEstablishing Trust in AI: Blockchain
46mCan you trust your AI?