AI Governance Skills Framework: Who is an AI governance professional?

As organizations increasingly adopt artificial intelligence, having trained professionals responsible for its safe and responsible implementation is more important than ever. But who are these professionals? What types of work are they tasked with? And what types of skills and knowledge are they expected to have? Do all organizations need an AI governance manager?

These are the types of questions that have driven our research into defining what an AI governance professional is. The outcome is our new AI Governance Skills Framework. The AIGSF is our way to define and document the AI governance workforce.  

As it turns out, the work of AI governance is not just one thing. Many have called AI governance a team sport, and our research validated that analogy. The work is not limited to one person or role; it is spread across a variety of roles and job functions.

Unlike a sports team with clearly defined roles and positions, AI governance is an evolving arena where the rules of play are still in flux. However, the field is beginning to coalesce around distinct types of AI governance work, including the types of work being tasked and the knowledge, skills and experience to perform AI governance activities.  

While these roles may be labeled and structured differently across various organizations, the basic tasks and professional responsibilities of AI governance are becoming increasingly apparent. This is why now is an ideal time to develop the AIGSF to better define these efforts.

Objective

As a professional association that dedicates significant effort to understanding the education and training needs of digital governance professionals, one major goal in developing the framework was to create a classification system and common language for discussing AI governance work. This classification provides a baseline for our research, training and engagement efforts. It also helps us understand where AI governance activities are situated within organizations.  

Publishing this classification gives those across the AI governance ecosystem to have a common terminology and shared understanding of the work, facilitating the exchange of best practices and fostering professional connections.  

The work of identifying not just the roles but also their main tasks, skills and knowledge helps the AI governance workforce evolve and grow. Research findings indicate many job profiles have overlapping responsibilities. Product, legal, engineering, and policy teams all play a role in AI governance, but without clear delineations of roles and responsibilities, siloed efforts can lead to redundant and misaligned work.  

Finally, defining these profiles will allow us to better understand the education and training needs required to advance the AI governance workforce. In future versions, we hope to map these profiles to a standard like the EU e-Competence Framework and/or education opportunities. However, market maturation is required before this can take place.

Methodology

To understand the scope of work individuals tasked with AI governance, we looked at responses from the IAPP's annual surveys from 2023-2026. These job and salary surveys helped us understand what work was being tasked within organizations. They provided insight into whether or not AI governance tasks were being added to existing positions, or if these were net new roles. Finally, it allowed us to understand how these various roles interact with others, and where they are positioned within the organization.  

Additionally, research from the UK Government, TechUK, Ada Lovelace Institute, Center for Democracy and Technology and Partnership on AI were instrumental in understanding the scope of responsibilities various AI governance professionals are tasked with.  

From January to July 2026, we collected public job profiles that explicitly called for AI governance work. These job profiles were taken from multiple regions, domains and industries to try and provide the clearest picture of the breadth of AI governance work.

Based on the analysis of this combined information, we categorized the work into eight types: policy, governance, technical, product, ethics, risk, legal and assurance. Following the categorization, research and public job profiles were mapped to a standardized job template.  

A note on the template: As part of the research for the AIGSF, we looked at other job classification work including the World Anti-Doping Agency's professional standards and role descriptors. Inspired by the work that ENISA, the European Agency for Cyber Security, has done to classify cybersecurity jobs with their European Cybersecurity Skills Framework, the AIGSF adopts the same template to describe cybersecurity profiles.  

Throughout the design and development process, our AI Governance Advisory Board provided valuable input and feedback, which ultimately led to the current version of the AIGSF. In addition to the advisory board members, input was provided by additional subject matter experts including those at WADA, ENISA, VDE and companies who have developed robust AI governance programs. Each of these contributors helped to shape and develop the AIGSF.

Categorization

Based on feedback from advisors and subject matter experts, the AIGSF is divided into two main sections: Part I: Executive Functions and Part II: Professional Profiles. While the framework initially focused on the functions directly responsible for AI governance work, the experts pointed out that identifying executives responsible for oversight and ultimate accountability of AI governance is equally important.

Part I

For the executive functions in Part I, our research identified the executive knowledge and accountability attributes that support AI governance decision-making. Given the number of executive roles identified, these are grouped into strategic domains. The framework only includes knowledge and accountability requirements for executive roles as AI governance only represents one component of the work that these roles do. As AI governance functions mature, it is likely that the number of executives involved will be streamlined. However, at the moment and depending on how organizational leadership is structured, there are valid reasons to involve multiple leaders across the company in various aspects of AI governance.  

It is important to note that these leaders are not likely to be involved in all AI governance decision-making — only when it is relevant to their area of practice. 

Part II

The original information captured for each role in Part II was limited to tasks, skills and knowledge. After seeing ENISA's European Cybersecurity Skills Framework and speaking with their team, we realized the value in expanding our categories — in particular, the section called alternative titles. For an industry that is still nascent, the work seems to be more common than the label.  

Based on feedback from experts, there was also a desire to expand the types of profiles within each category to reflect the different versions of work including more senior and junior roles and various functions. It is important to note that while these are referred to as profiles, it is unlikely that any given organization will require every profile to be included within their AI governance teams. There are overlaps between the deliverables and tasks within several of these profiles. This reflects the current state of AI governance, where responsibilities are often distributed across a variety of roles and functions. As the field matures, these responsibilities are expected to become more standardized and streamlined.

Next steps

This work has been an attempt to identify and document AI governance roles and functions. Including and the types of people tasked with doing this work. Since this research began, roles and labels have evolved. For instance, it is likely that in the future words like safety will be included in one of the standard categories. For now, the framework sticks to team types that are already known within organizations as the work of digital governance is not new. Similarly, this work is not completely new. While the ways that we assess risk and ensure concepts of safety might evolve how we deliver governance for AI systems.

As such, this AIGSF will evolve, we are already anticipating what the next version will look like. However, having a common way to describe AI governance roles will be helpful to organizations who are seeking to build an AI governance team, or evolve current roles within their organization to better support AI adoption.

As has been identified in our 2025 AI governance professional report. These categories of work define discreet job profiles that we are seeing in the marketplace. However, it is unlikely that most organizations will have a different person(s) responsible for each of these profiles. It is more likely that there will be a single role responsible for two or more of these functions. The larger and more complex the organization is, and depending on the type of AI they develop or deploy there will be a dependency on different aspects of the skills profile.

At present, the AIGSF is industry and domain agnostic. One common piece of feedback received during the development phase of the AIGSF is that governance for AI developers could be different than AI deployers. This is something that we will seek to explore more in future versions of the AIGSF.

As the AI governance industry matures, so will the AIGSF. It is likely that we will see movement and simplification of these categories over time as AI governance becomes a better understood profession.