AI and machine learning models at Manulife are governed through our enterprise Model Risk Management (MRM) framework. MRM sets consistent standards for how models are designed, tested, approved, and monitored over time.

Before approval, models must undergo validation and reliability reviews proportionate to their risk, with the scope, depth, and frequency of assessments calibrated to the model’s materiality, complexity, and potential impact, and risk-based monitoring maintained once deployed.

MRM also helps embed ethical considerations early in AI development. Teams complete structured checks to identify ethical and customer risks.

These steps aim to ensure AI solutions are evaluated not only for performance, but for alignment with Manulife’s ethical standards, legal requirements, and commitment to customer and stakeholder trust.

MRM follows a three‑lines‑of‑defense approach, with defined responsibilities across model owners, independent risk teams, and internal audit. This seeks to ensure reliability and explainability are ongoing obligations, not one‑time reviews.

Responsible AI Training

To support clear and effective ownership, Manulife mandates Model Owner training, helping model owners understand their responsibilities for ethical use, appropriate controls, and ongoing oversight throughout a model’s lifecycle.

In addition, all colleagues have access to training on using AI responsibly and securely, covering appropriate use, data handling, and where human judgment is required.

Monitoring for drift and degradation

Once a model is deployed, monitoring continues on a risk-based schedule to identify drift, performance degradation, changes in underlying data, or other issues that could impact the model's reliability, fairness, safety, or compliance with requirements. Where monitoring identifies an issue, the model is reviewed and revalidated, adjusted, or withdrawn from use.

The result

AI models are deployed with clear documentation, appropriate oversight, and ongoing  monitoring. This supports reliable outcomes, understandable decisions, and clear accountability in line with Manulife’s Code of Business Conduct and Ethics and enterprise risk expectations.

Transparency in how AI appears to users

We apply enterprise brand and design standards when AI is presented across our interactions with stakeholders. These standards set guidelines for identifying meaningful AI-generated content and how AI features are labelled and visually represented, so that colleagues and customers can recognize when they are interacting with AI.

By optimizing how AI solutions are built, leveraging smaller models where appropriate, and partnering with values-aligned vendors, Manulife seeks to integrate sustainability and energy-efficiency considerations directly into the AI development lifecycle.

Manulife’s partnerships with leading vendors supports our commitment to a sustainable future. Through these partnerships, we:

• Select AI solutions optimized for cost, speed, and efficient resource use
• Optimize how computing resources are used during model training, tuning and inference
• Proactively monitor our infrastructure usage to help reduce overall energy consumption

Manulife maintains a dedicated AI Governance team, responsible for defining enterprise-wide standards, policies, and control requirements for AI. This team works in close partnership with legal, compliance, privacy, risk, and model owners. Our Responsible AI Principles are endorsed by executive management, including the Chief AI Officer and Chief Risk Officer with ongoing oversight through an AI Governance Council.

Strategic oversight is provided by the AI Governance Council, operating as a GLT-level Steering Committee. This council brings together senior leaders to:

• Set enterprise direction for AI adoption
• Prioritize high impact AI use cases
• Ensure accountability and consistency across business and functions

Model owners reside in the business and function teams and are accountable for making sure models align with business objectives and governance requirements by setting tolerance limits and performance monitoring thresholds.

This layered governance structure helps ensure safety, compliance, and ethical considerations are embedded throughout the AI lifecycle, while enabling AI solutions to operate within defined risk tolerances and adapt to evolving technologies, regulations, and stakeholder expectations.

Security and privacy

Protecting personal information and maintaining the security of our systems and data are fundamental to how we use AI. We are committed to handling personal information responsibly and in accordance with our Global Privacy Notice. Privacy considerations are incorporated throughout the lifecycle of AI solutions, where applicable, to help ensure personal information is handled and protected appropriately.

We apply safeguards designed to protect the systems, information, and access that support AI, manage security risks, and support the safe, secure, and responsible use of AI. AI solutions are developed and governed in accordance with Manulife's information security and cyber risk standards. Access to AI systems and information is limited to authorized individuals with appropriate responsibilities and for legitimate business purposes. To learn more, see our Global Information Security Policy.

Manulife invests in continuous learning to stay aligned with evolving AI technologies, governance practices, and regulatory expectations.

We recognize that generative AI is evolving rapidly, and our governance, controls, and practices are designed to evolve with it. This establishes a clear baseline while continuously learning from new risks, technologies, and regulatory developments.

Our senior AI leaders regularly participate in and speak at Responsible AI and AI governance conferences, contributing practical experience and engaging with peers on the safe and ethical use of AI.

We collaborate with leading academic institutions, including MIT, York University, the University of Waterloo, and the National University of Singapore, to stay closely connected to emerging research and applied best practices.

Manulife also supports public dialogue through initiatives such as the Hinton Lectures, which convene global experts to explore AI safety, ethics, and societal impact.