Research Integrity and Safety

Important considerations regarding your use of GenAI

Responsibilities

As is the case for complying with all university, sponsor, and other applicable policies and regulations, the responsibility for understanding the GenAI tools used in the conduct of research rests with the principal investigator. Research team members are responsible for understanding the GenAI tools they use in their research, best practices within the discipline, as well as complying with all university, sponsor, and other applicable policies and regulations. Research teams must confirm the accuracy of all GenAI-produced materials and verify and document data sources used. See the USG Artificial Intelligence Guidelines: A USG IT Handbook Companion Guide for guidance regarding accuracy verification.

Data/Intellectual Property Ownership

UGA owns data and other materials or intellectual property generated in the conduct of UGA research, or third parties may control it through a sponsored research agreement or other contract. Researchers who use Gen AI tools are responsible for understanding if they control rights to any data or other material input to an AI tool. Some AI tools claim rights to inputs and products of the tool use, whereas others do not. Researchers should be aware of the rights associated with any tools and potential conflicts with university, sponsor, and other policies and regulations.

Data Privacy and Protection

The UGA EITS Data Classification and Protection Standard classifies data into three types: Public Data, Internal Data, and Sensitive Data and describes their characteristics and risks and requirements in place with respect to their use in AI models or AI powered search engines.

Consideration of the Impact of AI Use

Researchers should consider the tradeoffs in their use of GenAI in their specific field or area of scholarship and weigh the benefits against the consequences, intended and unintended. Questions of safety, societal impact, and ethical issues associated with the direct outcomes of the research need to be considered, and a strategy to minimize and then manage associated risks should be developed and implemented.

Protection of Human Subjects

As is the case with all research involving human subjects, their protection is paramount. Researchers need to be aware of the potential risks and limitations of the use of GenAI including biased and inaccurate data, the need for secondary source validation, and privacy and confidentiality risks posed by introduction of even de-identified data to AI tools. Contact the Office of Research Integrity and Safety (ORIS) with any questions you may have.

Patent and Copyright Considerations

Inventors of patents need to be actual individuals under the U.S. law. Documentation of human contribution and disclosure of the nature of GenAI utilization are therefore essential for patent eligibility. Similarly, copyright can protect only material that is the product of human creativity and authors of copyrighted materials must be actual individuals, so clear delineation of human and GenAI contributions/use is essential.

Perspectives and Cultural Norms (Cornell University)

Cornell’s Generative AI in Academic Research: Perspectives & Cultural Norms (Dec 2023) summarizes the following individual and shared duties we all have as well-informed academic researchers and scholars:

Duty OF VERIFICATION

• DO verify the accuracy and validity of GenAI The responsibility for research accuracy remains with researchers.
• DO check for unintentional GenAI can produce verbatim copies of existing work, or more subtly, introduce ideas and results from other sources but provide incorrect or missing citations.

Duty OF DISCLOSURE

• DO keep documentation and provide disclosure of GenAI use in all aspects of a research process, in accordance with the principles of research reproducibility, research transparency, authorship and inventorship.

Duty of DISCRETION

• DO NOT assume that GenAI is GenAI systems run on training examples, and user input and behavior are a prime source. Even if organizations that provide GenAI tools do not currently claim to use data in this way, there is no guarantee that they may not in the future.
• DO NOT share confidential, sensitive, proprietary, and export-controlled information with publicly available GenAI tools.
• DO NOT assume that GenAI output is already considered part of the public domain (e.g., not legally encumbered by copyright). GenAI tools can “memorize” their training data and repeat it with a level of verbatim accuracy that violates copyright. Even material that is not copyrighted may produce liability for corporate partners in sponsored research, if it is derived from data generated by a competitor.