Mark Thirlwell, Global Digital Director at BSI.
“The buck stops here.” When Harry Truman popularized the phrase, he was speaking about presidential power and responsibility. Roughly 80 years on, we’re starting to see this idea of personal accountability applied to AI, with governance moving from sitting within information and communications technology (ICT) or strategy functions to being a board‑level and executive consideration.
This makes sense when you consider that every business function will now interact in some way with AI, whether that’s finance using automated payroll tools, HR using it to review candidate applications or machine learning being applied to enhance data analysis and information gathering.
As things stand, no single accountability model is dominating, although our August 2026 research revealed rapid growth of the chief AI officer (CAIO) title, with more than 20,000 CXO job titles on LinkedIn including “AI” or “artificial intelligence.” Several big-name firms have nailed their colors to the mast in this way. For example, Target and Deloitte have made headlines announcing CAIO appointments.
Whether it’s one CAIO or a select group overseeing AI usage, the prevailing question organizations need to consider is this: With whom does the buck actually stop? Does this new crop of CAIOs (or the boards, councils or CIOs given the responsibility) have genuine oversight and accountability, or, in reality, is it more fragmented?
For example, are decisions about AI in recruitment actually sitting with the chief people officer? Are commercial leads procuring AI sales tools without consultation? Is there a clear chain of command? To reframe Truman’s saying, it’s not just where the buck stops but also whether everyone knows where it does and is willing to adhere to this.
This comes against a backdrop of generally poor governance of AI, where it would appear guardrails need to catch up. When we assessed AI governance practices in October 2025, fewer than a quarter of surveyed organizations globally reported having an AI governance program (24%), and under half said use was controlled by formal processes. A year on, we might expect that to be closing, but not necessarily fast enough.
This is concerning in general, but it’s especially concerning if things go wrong. Already, we’re seeing increasing evidence of major AI failures, whether to do with AI models going rogue, data breaches or inaccuracies in information. The potential for reputational damage, financial risk or legal liability is huge, not to mention safety implications of flawed AI usage in healthcare, food production or security.
The hope is that AI maturity will drive governance maturity so that as firms move from experimentation to widespread operationalizing of AI use, they’ll concurrently establish more formal oversight.
That could mean a CAIO, but more crucially, it means senior leaders establishing and clearly articulating the AI strategy and tolerance for organizational risk. More than that, it means forming an AI accountability plan that suits the specific context and use case.
Building An Accountability Plan
Leaders don’t need to start from a blank page when building an AI accountability plan. Established frameworks such as ISO/IEC 42001 provide a structured approach for identifying where AI systems are being used in an organization, defining clear roles and responsibilities for AI oversight and assessing and managing AI-related risks. By taking a governance-led approach, organizations can embed accountability into the design, deployment and operation of AI systems rather than treating it as a compliance exercise.
Standards reflect internationally recognized best practice, developed through global consensus among experts from industry, government and other stakeholder groups across more than 50 countries. They give leaders a practical and credible foundation for building AI accountability programs that are scalable, risk-based and aligned with evolving regulatory and stakeholder expectations.
Bringing Others On The Journey
Who exactly should be involved in AI decision making depends on the organization’s size, structure and governance model. In any type of organization, however, it’s critical to establish clear lines of accountability from the onset as part of an effective AI governance framework.
Leaders should define roles and responsibilities for AI oversight, ensuring ownership of AI-related risks, decisions and outcomes is clearly assigned. This includes responsibility for assessing AI risks, documenting policies for AI use and data governance, monitoring AI system performance and impacts and implementing corrective actions when issues are identified.
Done well, this translates across the organization, helping functional leaders understand how this relates to their remit with clear roles, responsibilities and authorities assigned.
Keeping Pace With AI
One of the most challenging aspects of creating an AI accountability plan is maintaining effective oversight as AI technologies and use cases rapidly evolve. Like cybersecurity, AI governance can’t be a one-time exercise; it requires clear policies, procedures and accountability mechanisms that are consistently applied across the organization and regularly updated to address emerging risks, changing business requirements and new regulatory expectations.
Organizations can best achieve this by embedding accountability throughout their governance system and establishing collaboration across senior leadership, risk and compliance functions, technical teams, data owners and operational stakeholders. A cross-functional approach enables organizations to identify and manage AI risks as they emerge, demonstrate responsible oversight and maintain trust with customers, regulators and other stakeholders while ensuring governance remains aligned with the pace of technological change.
Business Realignment And Adaptability
How organizations use AI is going to engender profound business realignment. Even as the parameters change, the constant will be the need for AI accountability—for leaders to manage AI and determine where the buck stops.
Whether or not they have a CAIO in post, those on the front foot will already recognize the value of robust governance and chains of ownership. In contrast, companies that shout about appointing a CAIO but fail to prioritize operationalizing this approach will be opening themselves up to a range of vulnerabilities.
Now is the time for leaders to build their AI accountability approach. Whether appointing a CAIO is the answer or not, only your organization will know.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
