Guidelines for Select Tool Use - Individual Tool Page Links
Guidelines for Select Tools
The following guidelines are provided to ensure that State of Georgia agencies effectively and ethically use AI tools, including ChatGPT, WellSaid, and Tableau AI. These guidelines focus on transparency, accountability, fairness, security, and the responsible use of AI technologies.
General Guidelines Across All Tools
Human Oversight
AI tools should augment, not replace, human decision-making. Agencies must ensure that a responsible human is always in charge of interpreting and validating AI-generated outputs.
Training and Awareness
Staff using these tools must undergo proper training to understand their capabilities and limitations. Regular updates and re-certification should be mandated to keep users current with the latest developments and best practices in AI use.
Ethical Considerations
Agencies should adhere to ethical AI principles, ensuring that these tools are used in ways that respect human rights, promote transparency, and avoid any form of harm or discrimination.
Audit and Monitoring
Regular audits should be conducted to monitor the use of these AI tools. Agencies must establish governance bodies to ensure adherence to AI principles and address any unintended consequences or misuse.
General Prohibitions Across All Tools
Circumventing Ethical Guidelines
Agencies cannot bypass ethical standards by using these AI tools in ways that contradict established ethical AI principles, such as fairness, transparency, and accountability.
Discriminatory Practices
AI tools cannot be used to intentionally or unintentionally reinforce discriminatory practices, such as excluding specific demographic groups from services or creating biased content or analytics.
Automating Sensitive Processes Without Human Oversight
Agencies cannot fully automate any sensitive process, such as legal compliance, budgeting decisions, or citizen interaction, without robust human oversight and intervention.
Engaging in Surveillance or Profiling
AI tools cannot be used to track, monitor, or profile individuals without explicit legal authorization, consent, and clear justification. Misuse for surveillance purposes is strictly prohibited.
Misrepresentation of AI Capabilities
Agencies cannot mislead the public or stakeholders about the capabilities or limitations of AI tools. Claims that AI is making fully autonomous decisions or replacing human judgment in key areas are prohibited.