The Role of AI Governance in Value Creation
Image: Depositphotos
Deploying artificial intelligence (AI) has complex challenges concerning ethics, transparency, bias, and fairness. AI governance can mitigate these challenges.
What is AI governance? OECD has proposed that artificial intelligence (AI) governance refers to the comprehensive framework of policies, regulations, ethical guidelines, and processes designed to oversee the development, deployment, and utilization of artificial intelligence (AI) systems in a manner that is ethical, transparent, and aligned with societal values. According to IBM, artificial intelligence (AI) governance refers to the processes, standards and guardrails that help ensure AI systems and tools are safe and ethical. AI governance frameworks direct AI research, development and application to help ensure safety, fairness and respect for human rights.
The challenge is not so much about defining AI governance but creating an environment where the actual day-to-day practices ensure AI systems are trustworthy, reliable, and compliant. As AI is evolving rapidly – so rapidly that by the time companies debate and implement policies – they may already be outdated. Further, many risks can't easily be predicted.[i] Done well, AI governance can mitigate risk, ensure compliance.
What often receives less attention is that AI governance also has an important role in value creation and driving innovation. AI must produce results and the context within which AI governance is framed does indeed matter. Despite the abundance of rhetoric around high-level frameworks, the biggest hurdle remains how to make AI governance work in practice.[ii]
As improved performance for both customers and corporations is the principal desired output from the deployment of AI, the governance structure has significant impact. The governance framework should enable organizations to move from “proof of concept” to “proof of value as Seth Earley stressed in assessing how to get beyond the hype with Generative AI. The mental model of the management team is crucial in creating the right framework for AI governance. It can be problematic when companies attempt to deploy AI in the context of a traditional departmental governance structure. That’s because departments will be tempted to act in their own best interest – which is not necessarily the same as that of the company nor that of their customers. Departments may hoard data – and may not collaborate on data sharing and structure with other departments. Similarly, departmental management attention may focus on their own self interests when it comes to performance measurement in a traditional departmental paradigm. AI projects typically focus on narrow tasks or sub-processes when the AI program is structured according to departmental goals.
These problems disappear when a company’s leadership team decides to use a customer focused, process-based framework for its AI governance. In this case, both transparency and the drive to standardize data formats across departments is built in. Performance metrics are defined and monitored according to what matters to the company and its customers.
When AI governance is framed in a traditional context, management attention is focused on departmental performance. Data format and structure may vary by department and data is not shared easily across the organization. The focus in performance management is mostly on volume and cost factors and attention to variance to budget by department dominates. Cross functional collaboration is ad hoc.
Viewing AI in a traditional context of departments makes performance management extremely difficult. It also stands in the way of the type and scope of change management needed for AI implementation success.
On the other hand, creating an AI governance framework in the context of the end-to-end processes that create value allows for greater transparency and innovation. This means that error rates, cycle times, and non-value-added activities can all be reduced through deploying AI. When AI governance is framed in a process based context, management attention is focused on value creation for processes such as idea to launch, inquiry to order, order to delivery, complaint to resolution, etc. Data is standardized and shared across the organization. In performance management, the focus includes factors such as quality and timeliness (which matter to customers) as well as volume and cost. Cross-functional collaboration and attention to customers is built in.
Implementing AI governance at the company level requires consistent execution. Projects are launched with a view to value creation. A cross functional team would check models, use checklists, generate change logs, and use real-time dashboards to track model performance. Within a process-based AI governance framework, cross-functional teams would also conduct risk and impact assessments as well as training on AI ethics, bias mitigation, data handling, and governance policies.
A shift in management attention from what departments do to how processes create value can improve the measurable results from AI projects. The synergy between process management and AI has been recognized. It makes it easier to manage large quantities of high-quality data needed to train AI and that deployment is enabled when the underlying process is well defined.[iii] A process-based AI governance framework has an important role in value creation and driving innovation.
Andrew Spanyi is the Editor of Cognitive World and the Founder of Spanyi International, which provides expert services at the intersection of customer experience, process innovation, and digital technologies. He is also a member of the Cognitive World Think Tank on enterprise AI.