As artificial intelligence (AI) and machine learning (ML) rapidly gain momentum as commonplace organisational functions, concerns about the ethical and responsible use of AI technology have also begun to surface. To help organisations build a framework for more transparent and unbiased AI, global IT research and advisory firm Info-Tech Research Group has published its new industry blueprint titled AI Governance.
Artificial intelligence is defined as a combination of technologies that can include machine learning. These AI systems perform tasks that mimic human intelligence, such as learning from experience and problem solving. Most importantly, AI can make its own decisions without human intervention. Machine learning is an AI process in which systems learn from experience without explicit instructions. Instead, patterns are “learned” and analysed from data to make predictions based on past behavior and learned patterns.
Often used to support or replace human decision making, the use of AI and ML offers several potential benefits, including enhanced customer experience, improved operational efficiency, and automated business processes. While this makes incorporating AI and ML functions into routine organisational operations appealing, doing so must be approached with clear expectations and concise governance procedures in place.
“Since ML and AI technologies are constantly evolving, so too must AI governance and risk management frameworks to verify that the appropriate safeguards and controls are in place,” says Irina Sedenko, advisory director at Info-Tech Research Group. “To ensure responsible, transparent, and ethical AI systems, organisations need to review existing risk control frameworks and update them to include AI risk management and impact assessment frameworks and processes.”
Based on the timely research included in the AI Governance blueprint from Info-Tech, the firm recommends that organisations consider the following key components for an effective AI governance framework:
- Monitoring – Monitoring compliance and risk of AI/ML systems or models in production.
- Organisation – Structure, roles, and responsibilities of the AI governance organisation.
- Operating Model – How AI governance operates and works with other organisational structures to deliver value.
- Risk & Compliance – Alignment with corporate risk management and ensuring compliance with regulations and assessment frameworks.
- Policies/Procedures/Standards – Policies and procedures to support the implementation of AI governance.
- Model Governance – Accountability and traceability for AI/ML models.
- Tools & Technologies – Tools and technologies to support the AI governance framework implementation.
There is no one-size-fits-all AI governance structure. As such, Info-Tech encourages organisations to identify roles and responsibilities at strategic, tactical, and operational levels; establish an AI governance council; and identify all groups supporting AI initiatives. The organisation’s maturity, size, and enterprise governance structure will influence the AI governance structure.Building an effective AI governance framework and program will help organisations:
- Define accountability and responsibility for AI.
- Define the AI risk management framework.
- Support the ethical, transparent, and fair use of AI.
- Define a framework to support ML/AI model governance.
The AI governance framework can also be used to define a set of metrics and key performance indicators (KPIs) that can measure the success of the framework implementation.