Jag Randhawa

Artificial intelligence (AI) is an unstoppable force that will transform all industries. Just like refrigeration technology transformed the food industry by removing barriers of localization, seasons, spoilage and safety, AI is going to transform all industries around how decisions are made, products are developed, services are delivered, and customers are engaged. Businesses that adopt AI will see unparalleled opportunities, growth, efficiencies, and innovation. However, with great power comes great responsibility. AI cannot be treated as a “black box” or something relegated to geeks in the windowless room. AI will enable businesses to operate and make decisions at unprecedented scale. When operating at such a scale, a seemingly trivial decision or negligence can put the company at serious risk.

Effective AI governance is crucial to ensure that AI systems are used ethically, safely, and in compliance with regulations. AI Governance should be a Board level topic and deserve the same level attention as the Audit Committee. One of emerging assurance services is the independent AI Audit, AI Governance Committee should engage an independent AI Auditor similar to how the Audit Committee engages the independent CPA financial auditor.

This article identifies key components of AI governance for business executives and the Board of Directors. According to NACD Board Governance Framework, the Board has fiduciary responsibility to oversee company strategy and risk. AI will be, and should be, part of every company’s strategy and AI also carries many inherent risks. It is imperative for the business leaders and the Board to establish AI Governance framework, policies and oversight.

What is AI Governance?

AI governance refers to the framework of policies, practices, and procedures that guide the development, deployment, and use of AI systems within an organization. The goal is to ensure that AI technologies are aligned with organizational objectives, ethical principles, legal requirements, and societal values.

AI governance should start with including Artificial Intelligence in the Enterprise Risk Management (ERM) framework. ERM scorecard for AI should include 1) ethical use of AI around how the decisions made through machine models impact customers, employees, stakeholders and society. 2) Compliance and regulatory requirements. Even though are no definitive guidelines published around AI governance but be careful how existing regulations like GDPR and CCPA may in principle be regulating use of data and information sharing. 3) Risk Management practice that reviews how use of available data may introduce bias or selection risks. And lastly, 4) Reputation risk around misuse, mistakes and trustworthiness among customers, partners, employees and stakeholders.

AI Governess Framework

Ethical Guidelines: Establish clear ethical principles for AI use that address fairness, transparency, accountability, and privacy. For example, if a company uses AI to personalize online shopping experiences, does the company disclose how their data is used to better serve them and if they have the choice to opt out of this practice? Does the company have guidelines to prohibit the use of AI for manipulative pricing or discriminatory practices.

Data Management: Ensure data used for AI is accurate, secure, and handled in compliance with existing and emerging regulations. For example, if a financial institution employs AI for credit scoring, has the company implemented strict data management policies to ensure customer data is anonymized and encrypted. How does the company ensure that data used does not introduce selection bias by disproportionality representing certain segment of customers, which could lead to bias and unfair decisions.

Bias Mitigation: Implement measures to identify and reduce biases in AI models, ensuring fair and unbiased outcomes. If a healthcare provider uses AI to diagnose medical conditions, how frequently do they audit AI results for biases and composition of patient data to include diverse datasets representing various demographic groups and populations.

Accountability Structures: Define roles and responsibilities for AI oversight that include members with diverse and broad perspectives. The committee includes members from diverse backgrounds, including ethics experts, legal advisors, and technologists, ensuring a holistic approach to AI governance.

Transparency and Explainability: Ensure AI decisions can be understood and explained to regulatory bodies, stakeholders, enhancing trust and accountability. For example, an insurance company uses AI to assess claims, they should have an explainable framework that they can show to regulators and customers why the company decided to settle or fight a claim or offer certain damages value to the claimant.

Continuous Monitoring: Regularly review AI systems for performance, compliance, and ethical considerations. When companies embark on AI journey, the first thing they realize is the need for more data. Sometimes it is internal data and other times it comes from external partnerships forged to gain insight into consumer behavior. As these data sources are added or AI models are optimized for accuracy, they have the potential of introducing or surfacing issues that were not previously discussed. It is critical that both management and the Board regularly discuss the use of data, model changes, recommendations, and risks associated with any of these.

Businesses have unique needs, and their value propositions are equally unique. It is the responsibility of the management to identify how artificial intelligence will add value to its business model and customers, however both management and the Board have the responsibility for ethical, fair and transparent use of AI. By implementing robust AI governance frameworks, organizations can harness the full potential of AI. Business executives and the Board play a crucial role in championing AI governance, fostering a culture of responsibility, and setting the stage for trustworthy AI usage.

I am looking forward to hearing your thoughts on effective AI Governance.