RiskVersity's Top 10 Business Risks Series: AI Governance

For many organizations, artificial intelligence is rapidly becoming part of everyday operations. You would be hard pressed to find an organization that hasn’t explored some of the utilities of AI tools, from HR screening to customer service chat bots to data analysis. Despite the adoption of these technologies, or at least the curiosity that surrounds their potential, the governance of AI is hardly keeping pace.

Research from the World Economic Forum identifies AI as one of the fastest-rising business risks. According to the Global Risks Report, 44 percent of organizations say AI introduces significant governance and compliance risks. In fact, despite being a technology that is in its relative infancy, modern organizations are already seeing some of the significant impacts of these risks, including situations that have impacted some of the smartest organizations on the international including:

• Data leaks– an IBM report found that a significant majority of reported data breaches resulted from improper AI access.
• Biased decision-making – There are numerous examples of bias based on AI models, with examples ranging from automated preferences for male job applicants to biases against certain patients in healthcare systems based on race.
• Compliance violations – Think HIPAA, discriminatory practices in violation of EEOO, and more.
• Reputational harm – In 2025, Grok, the AI chatbot from Elon Musk’s empire, made a series of antisemitic posts to users, creating a huge vulnerability regarding xAI’s reputation with consumers as a trustworthy AI creator.

As seen in these examples and many more, AI is a tool—not the answer—and requires oversight and review to ensure it does not lead users down a path of no return. Take, for instance, this case out of Raleigh, NC. An acting attorney used AI to create a legal argument for the Department of Justice and failed to check the validity of the AI-generated information. The filing cited court cases that didn’t exist and quotations that were never said. The attorney immediately lost his job, and the situation prompted warnings from the lead U.S. attorney for the Eastern District of North Carolina that AI may “hallucinate.”

Real-world examples continue to emerge. Despite these risks, the reality is that many can be managed with proper AI governance. Responsible organizations that are ready to embrace the potential of AI must ensure they have proper guidelines, protocols, and AI oversight in place to answer questions about who owns AI decisions, how exceptions are handled, and how AI usage is reviewed over time. These questions can be addressed in formal processes and protocols, including:

• AI usage policies
• Data protection controls
• Human oversight
• Ethical standards and audit trails

At RiskVersity, we continue to think about the many ways that AI impacts modern organizations, including exploring both the benefits and drawbacks of AI in various applications, like workforce (read our blog ‘Robotics and the Displacement of Human Labor’). Unlike many organizations, our approach to AI governance is risk-first, not tech-first. This lens enables us to see where AI aligns—and creates friction—with compliance, insurance, human resources, operations, and more. We help organizations build AI governance frameworks that enable innovation while protecting data, compliance posture, and public trust—building strong organizations that are ready for the next phase of tech innovations.