State of Georgia
Generative AI Standards for State Governance
Generative AI is revolutionizing public service operations, and The State of Georgia remains committed to advancing the responsible use of emerging technologies as part of its broader strategy to promote efficiency, transparency, and public trust. In alignment with this mission, the Georgia Technology Authority (GTA) guides the development, implementation, and oversight of AI policies that shape the work of over 40,000 state employees. Building on last year's foundational policies, "SS-23-002: AI Responsible Use (Standard)" and "PS-23-001: Enterprise Responsible Use (Policy)," the Paragon Fellows have contributed extensively to updating and refining these AI standards to address emerging challenges in AI governance.
During the Fall 2024 term, the Paragon Policy Fellows supported GTA by conducting in-depth research and collaborative policy development. This work focused on identifying and addressing gaps in the existing guidelines, ensuring AI policies are adaptable and robust enough to accommodate rapid technological advancements. Our approach emphasized creating clear, actionable standards that can be implemented effectively across departments.
Through the efforts of the Fall 2024 Paragon Policy Fellows, the State of Georgia now possesses an updated framework that reflects the rapid evolution of Generative AI technologies. This framework emphasizes clarity, ethical practices, and a commitment to supporting state employees in navigating the complexities of AI in their roles.
This projected aimed to position Georgia as a leader in Generative AI governance by:
- Updating existing standards to address new challenges in AI applications.
- Supporting state employees through actionable guidelines and resources for responsible AI usage
- Strengthening GTA’s capacity to promote ethical AI practices across government operations
To achieve these objectives, our team followed a structured, iterative approach:
- Expanded the AI Glossary: Added 16 terms and refined definitions to improve accessibility and understanding of AI-related concepts.
- Resolved Document Feedback: Addressed over 25 comments on the current standards document to incorporate best practices.
- Developed Supporting Appendices: Proposed supplementary materials, such as best practices for maintaining AI query records and crafting unbiased prompts.
- Conducted Comparative Analysis: Reviewed policies from other states and municipalities to benchmark best practices.
Literature Review
Summaries of current efforts in AI education, regulation, applications, and governance, drawn from both private and public sectors.
AI Definitions: Placements Guide
Revised guidelines informed by feedback from the Georgia Technology Council, addressing transparency and accountability.
GenAI Policy/Guideline Tracker
A review of AI policies from 24 cities and states.
Use Case Table
Six key scenarios with permitted and prohibited uses of AI.
Query Writing and Response Analysis
Methods for maintaining query records suitable for audits and bias assessments.
Unbiased Queries Notes:
Tips for writing impartial and equitable prompts.
AI Tools Approval Process
Guidance on enforcing standards, engaging employees through an AI Ambassador program, and conducting collaborative workshops.
The updated Generative AI standards signify a crucial step forward in equipping Georgia’s state employees with the tools and knowledge needed to navigate AI’s complexities responsibly. These standards provide a roadmap for ethical AI use, emphasizing transparency, accountability, and adaptability.
Looking ahead, we recommend the following:
- Build a detailed and flexible AI policy framework that can adapt to rapid technological advancements and evolving use cases.
- Invest in ongoing education and training programs to help employees stay informed about AI developments and best practices.
- Establish clear and concise documentation to ensure that AI use across agencies remains consistent, transparent, and compliant with regulations.
- Create dedicated spaces for collaboration where employees can share AI innovations, challenges, and insights to build collective growth.
- Regularly update approved tool lists and use-case guidelines to reflect emerging trends and tools, ensuring relevance and effectiveness.
Nsisong Udosen (Project Lead)
Parsons School of Design
Abhay Sastry
Georgia Institute of Technology
Desiree B. Junfijiah
Widener University
Elena Ajayi
St. John's University
Noreen Ahmed
Stanford University
Sharanya Maddukuri
George Mason University
Shivam Saran
Emory University
Wynee Pintado
Barnard University
Zev Van Zanten
Duke University