Our audit objective was to determine whether USPTO had an effective governance structure and processes in place to manage its AI tools. To meet our objective, we tested two of the six AI tools USPTO had in use when we began our audit.
Overall, we found that USPTO has begun developing its AI workforce but should strengthen key organizational and system-level governance practices needed to effectively manage and oversee its AI tools. Specifically, USPTO:
- Has a governance structure that defines roles and responsibilities for key personnel, but should improve internal stakeholder involvement
- Should promote transparency on its AI tools to external stakeholders
- Does not have the specific, measurable objectives needed to define system success
- Did not trace requirements or technical specifications to system objectives
- Does not have an AI-specific risk management plan
Together, these weaknesses increase the risk that USPTO will develop unreliable, untrustworthy AI systems.