Banner artwork by Moor Studio / Shutterstock.com
Team exercises aren’t just for athletes, military pilots, and business groups. They can also be valuable tools for lawyers supervising AI. Maureen, a GC for a fast-growing software company, recently told me she conducts "Red Team vs. Blue Team" exercises to identify and mitigate AI-related risks.
“The Red Team acts as attackers trying to exploit vulnerabilities and raise legal concerns about our AI systems,” she said. “The Blue Team defends and reinforces the systems, helping us think critically about potential risks as we prepare for real-world scenarios.”
Maureen effectively guides team exercises, because she, like many lawyers, has long been accustomed to supervising others. Corporate lawyers often ensure that a company’s legal and business personnel adhere to legal, ethical, and industry standards. It naturally follows that GCs would also oversee those using AI systems. Taking on this oversight responsibility requires a robust supervisory framework that sets out the principles governing the use and improvement of your AI systems.
How do you approach the development of a comprehensive AI supervisory framework that works for your organization? Red Team vs. Blue Team exercises are just one of the several considerations. Here are four more:
1. A two-pronged approach to interdisciplinary collaborations.
Effective models for AI supervision should combine multiple perspectives from end-users and a variety of professionals inside and outside the organization.
The AI supervisory role fits GCs so well because lawyers often recognize the value of well-defined systems that promote accountability. Many naturally appreciate governance frameworks that emphasize transparency, compliance, and fairness.
GCs and legal teams must collaborate with technologists such as data scientists, AI engineers, and experts in relevant areas like ethics. Their input enhances your overall understanding of AI systems’ functionality and real-world implications. You gain much-needed clarity for envisioning well-rounded processes that address multiple aspects of AI governance.
Collaborate also with leaders and end-users in every business unit that uses AI; their daily workflows tend to be the most affected. A robust supervisory framework will include steps to incorporate feedback from actual users to fine-tune your AI models. Promoting cross-functional discussions now carves a path for future feedback. At the same time, brainstorming sessions often lead to eye-opening discoveries and help you proactively spot risks that may otherwise go overlooked.
2. Experiment with multiple AI tools and platforms.
Explore various AI-supported platforms and experiment with AI-driven features to learn how AI functions in practical contexts. Hands-on experience gives you a more thorough understanding of AI’s capabilities, risks, and limitations, facilitating more effective communication with technical team members.
Many free tools are available online. You may have access to more AI tools than you initially thought, as many technology vendors have incorporated generative AI tools into their recent offerings. You must inventory all the company’s AI tools to supervise them. During that process, you or your supervisory team members can explore their features to consider how existing regulations, particularly those related to data privacy and IP protections, apply. Firsthand knowledge enhances your ability to tailor your supervisory framework to your company’s specific needs, as you know how AI tools work in the real world.
3. Seek guidance from other industries.
Another reason the AI supervisory role fits GCs so well is that lawyers often recognize the value of well-defined systems that promote accountability. Many naturally appreciate governance frameworks that emphasize transparency, compliance, and fairness.
You can find potential governance frameworks in industries beyond your own. Try looking into how other sectors regulate and supervise AI. Explore case studies, best practices, and guidelines from the healthcare, finance, and transportation industries. Seeing how different fields manage their AI systems can help you craft novel approaches to your supervisory structure, including strategies for increasing awareness and training employees.
4. Monitor AI development with internal and external audits.
Regular internal audits play a significant role in achieving and maintaining the efficacy of an adaptable AI supervisory framework. Periodic assessments serve as guideposts, revealing new areas of concern as growth occurs and helping ensure your AI systems continuously align with the firm’s evolving objectives and compliance requirements.
Firsthand knowledge enhances your ability to tailor your supervisory framework to your company's specific needs, as you know how AI tools work in the real world.
In addition, explore the possibility of engaging with an independent auditor to scrutinize your supervisory framework and methods for maintaining compliance, performance, and ethical standards. Outside auditors can offer the latest expertise in the fast-moving field of AI and suggest useful approaches to supervising AI.
As AI reshapes the legal landscape, lawyers' responsibilities expand.
Developing a comprehensive AI supervisory framework helps you stay ahead of the curve and ensure your organization’s use of AI remains efficient, secure, and accountable. By maintaining oversight over AI usage and development, you directly address ethical considerations, enhance the quality of decision-making, and mitigate potential risks. Stepping up to the plate with a safety-first approach supports the right balance between innovation and responsible governance in the age of artificial intelligence.
