Use AI Responsibly

The Cybersecurity Governance Committee under the auspice of 麻豆社 Artificial Intelligence (AI) Innovation Committee has developed the following set of AI guidelines that align with industry recommendations and University System of Georgia (USG) cybersecurity recommendations. All 麻豆社 members should adhere to these guidelines when using AI services or applications that utilize any 麻豆社 technology resource (information technology, hardware, software, etc.).

For additional technology-use guidance, visit 麻豆社 Information Technology Policy webpage. 

AI Guidelines

  • Have a clear goal for learning or improving your business when using AI tools. This will help you determine if the AI’s answers and results fit your purpose. 
  • Involve those who will be impacted by AI use when deciding to use it in the classroom, for departmental or organizational purposes, or for research. This includes those who will create, use, manage, or be affected by the output of an artificial intelligence system.
  • Check if results provided by a service using AI are accurate and appropriate for all potential recipients of the information. Test common and uncommon prompts before any technology system becomes available for use by others and continue testing regularly after it is in use.  

Example: An AI tool asked to draft research bibliography citations should provide results that match the proper writing style handbook.

  • Protect sensitive information entered into the AI system appropriately and address any worries that people have about AI use. 

Example: Social security numbers and credit card information should not entered as part of a AI model prompt.

  • Evaluate the benefits and risks for the university, especially regarding data security and privacy, before purchase, during implementation, and after beginning use of any AI system; including that of new AI software, new AI features in existing software, and AI applications that work with other systems. 
  • Teach employees and students how to use AI systems properly and ethically to protect the privacy of sensitive data and to meet their business and educational goals.  

  • Train anyone using AI about the risks and benefits and how to spot bias in AI answers or analyses.   
  • Maintain formal ethics policies and codes of conduct for use of any technology tool, including artificial intelligence systems.  
  • Check whether the content you are using is protected by copyright and ensure you are not violating someone else's intellectual property rights by copying or using their work without permission.

Tip: AI models should not be trained using electronic books without the permission of the author or publisher.

  • Align AI use in any setting with existing laws, guidelines, institutional policies, or procedures. 
  • The use of AI systems that do not follow data security policies or provide inappropriate results should be paused so improvements can be made and to avoid harming the institution.
  • Whenever possible, data should be anonymized before inputting it into AI systems to remove any details that could identify people or expose sensitive institutional information.  

Example: Medical research using AI models should remove names and use other identification for patients whose medical information is used in the study.

  • Projects and services that use artificial intelligence should have documentation for deployment and testing that includes the development, implementation, and use of the system.   

Tip: Include the documentation of the artificial intelligence model use information within the project plan.

  • Research projects that use artificial intelligence should provide documentation detailing how artificial intelligence was used in the research.  

  •  Use artificial intelligence as an analysis tool and not as a substitute for human decision-making.  
  •  Define roles and responsibilities for implementing and maintaining artificial intelligence systems.