AI Use Disclosure

Guiding Frameworks

We approach AI use through guidelines rather than policy which reflects that this technology is evolving and, in that process, so is our understanding of it.

To that end, we align ourselves with University of Illinois System AI Guidelines (Administrative Information Technology Services, 2026). In practice, this means we use generative AI only with public data as an input. Public data is information intended for public release with no adverse impact if disclosed. We never load personal or sensitive data, such as student data, into generative AI tools.

Our Standards

We are also guided by the principle of Human-in-Power (Zheng et al., 2024), which centers context, ethics, and awareness of power imbalances in the use of AI tools. This means that all AI use at SMMC is supervised and no AI-generated content is published without human oversight and extensive review.

We are transparent about how we use AI. Current uses include reformatting educational materials, research synthesis with verified documents to explore trends, drafting and editing educational and social media content, and coding assistance. We do not use AI to generate decorative images, but we do use it to develop code to generate digital learning objects and data visualizations. We do not ask AI models to generate visualizations directly, as this makes it difficult to audit the process and detect hallucinations.

Our current uses for AI may evolve alongside these models. In all of these use cases, current and future, we always review and verify the resultant artifacts to ensure their accuracy.

We explicitly disclose when generative AI has been heavily used to create content, meaning more than 50% of the resulting artifact was produced by a model without human editing. We also note when core interactive elements, such as digital learning objects or visualizations embedded in our blog posts or website, rely heavily on AI.

Ethics in AI Use

As a unit dedicated to economic empowerment through financial literacy, we have a responsibility to acknowledge the ethical concerns that generative AI use entails.

For example, one of the reasons we explicitly do not use generative AI to produce images is because of how most models are often trained on the creative labor of visual artists without compensation. We acknowledge, however, that this is also true of the written material used to train textual models. Our decision to accept the ethical concern of using AI to generate written material and not visual material exemplifies the sort of messy decisions that must be made when employing AI as a tool.

Additionally, the widespread use of generative AI affects socioeconomic sustainability. Despite its potential for increasing productivity, the impacts of generative AI use on the environment and humans, particularly the most vulnerable communities, should not be ignored.

Five primary risks to society have been identified which include (1) lack of accountability, (2) lack of data privacy, (3) cybersecurity concerns, (4) unsafe systems, and (5) unintentional bias (U.S. Government Accountability Office, 2025). As a small unit, we cannot address all of these risks at large scale, but this disclosure along with ongoing thoughtful use of generative AI is a small step towards doing what we can. Additionally, we also proactively work to mitigate these risks through our educational outreach, such as by addressing misinformation and fraud stemming from increased use of generative AI and providing tools for others to mitigate these risks in their own lives.

Our Approach to Learning

AI tools are new, and best practices are still evolving. We are learning alongside the field and we acknowledge that mistakes may occur in this process. When they do, we will acknowledge them openly and adjust accordingly.