Preus Library's AI Literacy Guide

When we rely on Generative AI too much, we risk undermining our own ability to think. Overdependence on AI conditions us to expect quick, easy answers that we don't have to work hard to obtain or create. But remember... it's beneficial to think hard and struggle with difficult questions. Using our brains strengthens our brains. When we outsource our thinking to an external tool (such as AI), that process is called "cognitive offloading". 

Before using GenAI, consider:

• Do I plan to take the time to carefully interpret and evaluate what the AI generates for me?
• Is using AI for this task depriving me of a valuable opportunity to exercise my mental muscles?

Generative AI consumes vast quantities of energy--a single generated AI response costs significantly more energy than a web or database search. As widespread use of GenAI grows, so too does the burden on the environment. Many of these harmful impacts are inequitably distributed along geographic and socioeconomic lines. While companies frequently market new developments for mitigating GenAI's negative environmental impacts, there is still a long way to go before GenAI can be considered "green". 

Before using GenAI for a task, consider: 

• Is using GenAI for this particular task worth the environmental cost? 
• How can you be thoughtful about when (and how often) you choose to use GenAI?

Because GenAI models like ChatGPT are trained on huge datasets, they often inherit bias and stereotypes. These tools do not critically evaluate information to avoid racism, sexism, or other types of harm. When generated content is uncredited, there is no way (within the tool itself) for you to fact-check by identifying and investigating the direct source of the information.

Before using GenAI, consider:

• How do you plan to mitigate (by careful prompting) and identify (by evaluating the output) any bias or stereotypes in AI-generated content?
• What non-AI resources can you consult to make sure you do not accidentally perpetuate harmful bias or stereotypes?

As the popular saying goes: "If you aren't paying for it, you're the product". Many GenAI models--especially free ones--may use your data and content to train their models. Disreputable companies may use or sell your data for profit. Your private information could be leaked or disclosed without your consent. While these tools have Terms & Conditions, they can be updated at any time. Even when the option is available to "opt out" of data collection, you still need to be aware to take advantage of the protection.

Before using GenAI, consider:

• How do you feel about GenAI companies using people's private or personal information without consent?
• Does the GenAI tool you plan to use handle privacy concerns well? Can you configure the settings to protect yourself and opt out of allowing the AI tool to collect and use your data? 

GenAI models are trained on vast quantities of scraped data--usually without attribution. Large language models (LLMs) like ChatGPT and Gemini generate their responses using content created by humans, but do not provide credit or compensation to the original creators. GenAI is disrupting copyright protections and and intellectual property, leading to concerns among artists, authors, and creators and multiple ongoing legal cases.

When deciding whether to use GenAI, consider:

• Do you think that the way that GenAI scrapes content without attribution should be considered stealing?
• Is it possible to use GenAI ethically (no matter how diligently you cite your AI usage) if it generates content using uncredited sources?

GenAI can be easily used to create fake sources, images, and videos. People have been spreading photoshopped images and "deepfakes" for ages--for malicious purposes, disinformation campaigns, or pranks--but GenAI makes it much simpler to do so. As the tools improve, the fakes are harder to spot. 

Before using GenAI, consider:

• How can you prepare yourself to identify misinformation (incorrect) and disinformation (deliberately misleading)?
• What can you do in your life to help foster a culture of critical thinking and ethical use of technology?

Cheating has always been a concern in higher education, but Generative AI is exacerbating the situation considerably. It's easier than ever for students to use AI to bypass difficult work, even though the process of struggling and doing that work would be beneficial for learning. Students receive many mixed messages about whether or not they are "allowed" to use AI. The line between acceptable and unacceptable AI use is often blurry. When the use of AI isn't properly acknowledged or cited, students and researchers submit AI-generated content as if it were their own unassisted work. There are AI-detection tools on the market, but their reliability and accuracy are poor.

Before using GenAI, consider:

• Students: Do you understand your professor's policy about using AI? If you use AI, how will you acknowledge and cite it? Will your own thoughts and personality be reflected in the final product?
• Faculty: How can you design assignments to encourage students to utilize their own thinking and critical voice? How can you emphasize process over product?

While the rise of GenAI has massive implications for the job market, it also impacts other aspects of human labor. We often hear concerns about AI replacing jobs, but GenAI actually depends on a hidden human workforce. To manage the large scale of their AI products, companies may rely on a underpaid and exploited labor, often from the Global South. 

Before using GenAI, consider:

• Do you know if the company who operates the GenAI tool uses exploited human labor?
• Is using GenAI for this particular task worth the risk of a hidden human cost?