The Data Exchange with Ben Lorica 2025 AI Governance Survey
Aug 28, 2025
David Talby, an expert in AI governance at John Snow Labs and Pacific AI, shares insights from the 2025 AI Governance Survey involving over 350 industry professionals. He reveals that despite 30% of organizations having AI models in production, there are alarming gaps in risk management practices. The discussion spans AI adoption trends, emphasizing the dual roles of developers and deployers. Talby also stresses the importance of formal policies and unified platforms for effective governance and collaboration in navigating the evolving regulatory landscape.
AI Snips
Chapters
Transcript
Episode notes
Deployers Are Also Developers
- Most teams act as both deployers and developers because post-training work (prompting, guardrails, customization) dominates development needs.
- Organizations rarely pre-train models from scratch and focus on downstream tuning and integration instead.
Prioritize Post-Training Work
- Focus on mastering post-training (instruction tuning, prompt engineering, grounding) rather than pre-training models from scratch.
- Optimize your platform and pipelines for rapid model swaps and downstream task improvements.
Grounding Outperforms Fine-Tuning Often
- Grounding with context, metadata and structured data often yields bigger gains than fine-tuning.
- Grounding is easier to update in real time and avoids heavy GPU costs for fine-tuning.