The advent of Gen AI has ushered in new opportunities and the need for new governance. Learn how Delta Dental of Michigan, Ohio, and Indiana updated their 2020 DGPO Best Practice Award-winning DG Program to include AI Governance.

Planned topics include:

• How we've approached the use and governance of AI across the organization
• An overview of the AI Guardrails we've created
• Examples of AI Use Cases we've approved (and perhaps some we've denied)

The adoption of artificial intelligence (AI) in radiology is accelerating, offering powerful solutions to address the growing volume and efficiency challenges faced by radiologists. While AI demonstrates significant value in automating repetitive tasks and streamlining workflows, concerns around accuracy and medicolegal liability remain barriers to widespread trust and adoption. To bridge this gap, the American College of Radiology (ACR) has pioneered governance frameworks and registries that promote safe, clinically effective AI integration. Central to this effort is Assess-AI, the first national registry for monitoring AI algorithm performance using real-world clinical data. Leveraging large language models (LLMs), Assess-AI extracts structured insights from radiology reports to generate “surrogate ground truth” for continuous performance benchmarking and national comparisons. This presentation explores the governance strategies, safeguards, and scalable monitoring methods that ensure AI adoption remains clinically relevant, safe, and trusted—empowering radiologists to focus on high-impact decision-making while advancing patient-centered care.

The push towards AI-enabled transformation means that organizations can both amplify existing risks, as well as incur new kinds of risks. Understanding what those risks are and how to mitigate them is critical for developing and deploying AI responsibly and sustainably.

‍

In this session, we’ll cover a range of organizational risks - from reputational to financial and technical - explaining how AI is entangled with each of them. We’ll use real-world examples to illustrate how organizations are experiencing these types of risks and, more importantly, what you can do to mitigate them. Taking control of AI risk is an important part of enabling a responsible AI culture. We’ll show you some practical ways to thoughtfully and intentionally apply "due diligence” to your AI development and deployment decisions.

AI is shifting from single large models to swarms of autonomous agents—multiplying risks and blurring accountability. This session explores how enterprises can balance centralized and federated governance, embed observability, and operationalize policy as code—building adaptive frameworks that safeguard trust while enabling speed and innovation.

Is adding AI to your environment a software purchase? Or is it more like hiring an employee?  With the rise of agentic and autonomous AI, issues that normally apply to humans may increasingly apply. These include:

‍

• Background checks (what entities have had the ability to modify your AI? What known risks does it have?)
• Job descriptions (what functions will be in scope and out of scope for this AI?)
• Management responsibility (who is, and is not, permitted to operate this AI?)
• Performance metrics and expectations (how will the efficacy of this AI be measured?)

Add to that questions such as "what happens if your AI creates a hostile work environment for your human employees?" and it becomes apparent that AI is more than "just another software package". This session delves into the extent to which IT, data management, procurement, and HR should be working together to manage the importation of AI into your environment.

As AI agents evolve from simple task automation to autonomous decision-makers, governance becomes a strategic imperative—not just a compliance checkbox. This session explores the emerging landscape of AI agent governance, drawing on real-world engagements and Sogeti’s infrastructure-first methodology. We’ll examine how organizations can design agentic systems with built-in observability, auditability, and policy-as-code guardrails to mitigate risks like hallucinations, bias, and uncontrolled autonomy. Attendees will learn how to assess agent readiness, define mission boundaries, and implement governance frameworks that scale across multi-agent workflows. The talk will also cover the trade-offs between control and autonomy, and how to align agent behavior with enterprise values and regulatory requirements. Whether you're deploying copilots, research agents, or workflow orchestrators, this session will equip you with the strategic and technical insights needed to govern AI agents responsibly and effectively.