Generative AI at the River Campus Libraries

Users have to understand that currently there may be no purely ethical use of generative artificial intelligence.  GenAI tools are built on the mass, indiscriminate consumption of data; an opaque training system; and a general disregard for copyright and intellectual property.  There are numerous other ethical issues inherent with the technology [link to resources under Further Reading] that further complicate the struggle to ethically interact with GenAI.

Here are some general guidelines towards practical, ethical use of GenAI:

1. Keep humanity in the loop.  Keep an eye on the GenAI output at all times.  Never take what's generated for granted and be sure to put your own mark on the output.
2. Understand that bias is baked into the system.  GenAI is built from data scraped across the internet, and will reflect the generalized biases of that data.  Using GenAI irresponsibly can perpetuate existing prejudices and biases, both obvious and subtle.
3. Prioritize transparent programs.  Some companies have taken steps towards creating more ethical GenAI programs. Work with these tools over the less ethical options to help promote future, more transparent endeavors.
4. Familiarize yourself with the tools.  Understand how the tools operate to better inform your interaction with the technology.
5. Insist on stronger ethical standards in GenAI.  Get involved with your field's professional organizations, identify how ethical issues may impact your field, and recommend stronger regulations.
6. Consider when (and when not) to use GenAI.  Use the right tool for the right project; don't depend on GenAI to do the work for you and use it thoughtlessly.  There is an environmental impact to using GenAI, so use your time and prompts wisely.

Some key best practices to keep in mind when working with GenAI:

• Never add sensitive personal information.  Many of these tools input scrape:  taking user inputs to train their tool.  Once this data is added it cannot be removed.  Personal identification information (PII) such as your social security number, driver's license, pin numbers, phone numbers, etc. shouldn't be added, nor should any protected health information (PHI) if you are working with medical data. 
• Never add proprietary data.  Similar to sensitive personal information, avoid adding proprietary information into these tools.  As per the NSF, any proposal data shared to GenAI programs outside of their firewall is considered public domain.  Losing control of your data can have serious repercussions, including losing funding.
• Document and be transparent about your use.  If you use GenAI, consistently record when and how it was used.  Describe and disclose how it was used in any final writing assignments.
• Check your output.  GenAI often generates false information ("hallucinations").  Examine your output and ensure there are no hallucinated references, statements, and other fabrications.