In September, Mesh Research, including the entire Humanities Commons team, met for a one-week, in-person retreat at Michigan State University (more on our retreat to come!).
At the retreat, we spent some time thinking and discussing our use of Generative AI (GAI). GAI tools like ChatGPT, Midjourney, and Bard are increasingly part of a multitude of tasks in our daily lives. We want to let you know how we are currently using generative AI (GAI) in our workflows and how we are thinking about it for the future. As the Commons is sponsored, hosted, and developed at Michigan State University, we’ve taken our initial guidance from the MSU Interim Guidance on Data Uses and Risks of Generative AI. Currently we are not using GAI in the creation of content (images or written) or user experience data collection. We do use GAI to assist in authoring code, using tools such as GitHub Copilot, though always interactively with a human developer. In the spirit of our value of experimentation, we are also starting to investigate the use of GAI, such as ChatGPT, for data analysis.
Whenever possible, we will use tools or versions of tools that do not train on or learn from our usage. You can always count on us to have a human in the loop; any code or content generated by AI is reviewed by a member of the team before its use. We do not intend to use GAI to outsource our work, but to enhance it.
We will continue to discuss GAI in our meetings and revisit tools and techniques at least quarterly. When and if we do move forward with GAI in other parts of our operations, we are committed to letting you know here on the blog and in the newsletter. We’d also love to hear from you about ways you would like to see HC engage with GAI in the future!