AI is transforming businesses and industries by automating decisions. The implications of this are profound. Replacing human decision making with machine “intelligence” can drive enormous value, but also introduces entirely new types of risks for the enterprise. Considering the significant potential advantages and associated liabilities stemming from the widespread use of AI, organizations globally are developing and implementing AI Governance programs.
Many organizations start with a broad definition of AI Governance, for example:
AI governance is the ability to direct, manage and monitor the AI activities of an organization.
The broad definition is appropriate: AI Governance should take account of the full impact of AI activities on an enterprise. In practice however, these same organizations may construe AI Governance more narrowly, with an emphasis on mitigating risks related to regulatory compliance or concerns around ethics, bias, fairness and transparency:
The objective of AI governance is to deliver transparent and ethical AI to establish accountability, responsibility and oversight.
This narrower view of AI Governance misses the fact that AI initiatives address defined business objectives that originate within functional organizations and business units. The success of these initiatives is measured in terms of their business contributions – e.g. increased revenue, lower costs, improved quality, increased customer satisfaction, better employee retention, etc. – as well as their risks and liabilities. As such, AI Governance programs that focus solely on AI liabilities and risks ignore consideration of the business contributions, costs and ROI for AI initiatives and therefore miss the value side of the equation.
Get the whitepaper!
