Building AI foundations

This guidance is for courts that are just beginning to consider AI. These courts should start their exploration of AI by building foundations. Refer to our "Understanding your court's AI readiness" guide to learn about and navigate the entire AI maturity spectrum. 

Governance

Governance structures must be in place to ensure ethical, effective, and equitable implementation of AI. A well-defined governance framework provides a structured approach to managing and identifying risks, setting standards, and maintaining public trust. Effective AI governance includes clear policies, safeguards against unintended consequences, and mechanisms for ongoing oversight.

Governance can take various forms, including an AI governance committee to oversee adoption, ensure compliance with legal and ethical standards, and address emerging challenges proactively. However, governance also extends to personnel training and stakeholder engagement, ensuring that those interacting with AI understand its capabilities and limitations.

Set up your governance structure.

Guiding principles

Documenting guiding principles for AI usage and implementation is essential to ensure that AI technologies are developed and deployed responsibly, ethically, and effectively. Providing decision makers with a set of clear guiding principles to look to at each stage helps to prevent misuse, mitigate risks, and maintain public trust.

A court's guiding principles statement should be a relatively short document (often, one to two pages), outlining the high-level values that will shape its AI-related decisions. It should articulate the commitments that court is making and the principles to which all court applications of AI must adhere.

Suggested core principles include: transparency, accountability, fairness, and human-centered design.

Discover how guiding principles aid decision-making.

Internal use policy

Internal use policies can provide guardrails that promote safe AI use by court personnel during the transition period while the court establishes its AI governance infrastructure and longer-term strategy. An initial policy should not aim to cover all AI-related issues or potential AI use scenarios. Rather, the focus should be on identifying and mitigating the most immediate risks that may be involved in court personnel using AI for court business. The policy should make clear to court personnel which AI practices are allowed or prohibited in the short term.

So far, courts have adopted two types of broad policies: requirement of court approval for each AI use, and guidelines for AI use.

Common policy provisions include:

• Statement of purpose
• Policy scope
• Definitions
• Statement on human oversight and responsibility
• Allowed and prohibited AI uses
• Response to a breach
• Approval authority
• Training
• Policy updates

Read excerpts from internal use policies.

Data governance assessment

Data governance is a framework encompassing the people, policies, processes, and technology that ensure high-quality data, data management, and data security. The quality of a court's data determines what kinds of AI innovations are possible and plays a critical role in successful implementation.

Courts should be equipped to:

• Understand where AI technologies are most needed, will have the greatest benefits, and will do the least harm.
• Identify specific court operations and tasks to enhance with AI and which tasks to leave to humans.
• Develop better-performing AI tools.
• Measure the performance of AI tools to ensure they meet the courts' needs before launch and make adjustments and improvements as needed.
• Make sound, data-driven decisions about court policies and practices.
• Share technology and knowledge across jurisdictions, coordinate with justice partners, and learn from other courts.

As courts begin to explore potential new AI solutions, it is vital that they assess and improve their broader data governance practices.

 The following are some initial data governance improvement recommendations:

• Improve data quality.
• Standardize data across jurisdictions.
• Build data literacy.

Evaluate your data governance.

Literacy strategy

AI literacy refers to the knowledge, attitudes, and skills needed to effectively interact with, critically evaluate, and responsibly use artificial intelligence systems. By developing a plan for building and maintaining AI literacy in the workforce, courts can ensure that all personnel are equipped to navigate an increasingly AI-integrated workplace with confidence and discernment. 

Recommended steps for developing a strategy:

1. Define AI literacy.
2. Delineate literacy milestones.
3. Update recruitment practices.
4. Identify education and training needs.
5. Update personnel evaluation processes.

Build your literacy strategy.
 

See the full "Building Foundations" section