Details
AI language models are now processing vast amounts of corporate data, delivering insights faster than ever. But here’s the key point: You can’t trust an LLM. However, you CAN trust AI systems using LLMs when they’re guided by a foundation of strong governance and accountability.
Join EDM Council and data.world for our upcoming webinar where we’ll explore how to build trustworthy systems out of untrustworthy components. Large Language Models (LLMs) can be powerful tools when harnessed effectively, but their unpredictability can create challenges—much like musicians needing a skilled conductor to bring everything into harmony. We’ll discuss how organizations can implement governance and data catalog frameworks that act as the ‘conductor,’ ensuring AI systems work transparently, reliably, and in compliance with regulations.
In this session, industry experts will share insights on building AI systems that integrate LLMs, emphasizing the need for governance, data lineage, and quality control to maintain trust in these systems. We’ll delve into real-world examples and case studies, demonstrating practical approaches to managing LLMs while maintaining control and accountability.
Here’s what we’ll cover:
- The challenges of integrating unpredictable LLMs into enterprise data workflows
- How data catalogs provide the governance, lineage, and quality control essential for trustworthy AI systems
- A practical demonstration of building trust in an LLM-based system for regulatory compliance
- Strategies for ensuring data integrity and transparency in AI-driven decision-making processes
Through diverse case studies, this webinar will provide a deeper understanding of how organizations across industries are navigating the complexities of AI-powered enterprise data management. Whether you’re a CTO looking to deploy advanced AI solutions or a CFO concerned about regulatory risks, you’ll walk away with actionable strategies for fostering trust and governance in AI systems.
This is an opportunity to gain insights into the intersection of AI and data governance, learning from real-world challenges and solutions that are shaping the future of enterprise data management.
Speakers
Post-event summary
The webinar titled “Ensuring Trust in the AI Era: Integrating LLMs with Data Catalogs,” hosted by EDM Council and data.world, explored the intersection of AI and data management, emphasizing the need for trust and governance when implementing AI solutions in organizations. The conversation was led by experts in the industry:
- Bryon Jacob, CTO, Co-founder, data.world
- Dean Allemang, Principal Solutions Architect, data.world
- Moderator: Eric. Bigelsen, Head of Strategic Business Development, EDM Council
The webinar highlighted that AI, particularly large language models (LLMs), must be combined with robust data governance to provide trustworthy, explainable insights. Bryon explained, “The path to using LLMs on data isn’t to train the LLM on your data and take the answers that you get back from that; the answer is to couple the LLM with a deterministic system that has structured knowledge.” This approach ensures that AI-generated responses are traceable back to governed data sources, avoiding common pitfalls like hallucinations and misleading conclusions.
Dean demonstrated a practical scenario involving Reg. W compliance, showcasing how a knowledge-driven system can spot governance gaps and automate problem identification. By integrating LLMs with data catalogs, the AI system could provide actionable insights while enabling human oversight through clear audit trails.
The session also addressed regulatory developments like the European AI Act and the U.S. NIST AI Risk Management Framework, which emphasize the importance of governance, cataloging, and provenance. With AI becoming more accessible across organizations, the speakers urged the need for data and AI literacy, underscoring EDM Council’s role in helping professionals adapt to these shifts.
In closing, the speakers reiterated that while AI holds transformative potential, its successful deployment depends on strong data management foundations that support transparency, compliance, and reliability.