Artificial Intelligence (AI) - A Resource Guide

• AI and Environmental Challenges 
  Flanagan, M. (2024). AI and environmental challenges. Environmental Innovations Initiative, University of Pennsylvania.

• If A Hammer Was Like AI 
  Axbom, P. (2023, August 22). If a hammer was like AI. UX Magazine.

Misinformation

Generative AI tools can help users brainstorm ideas, organize information, plan scholarly discussions, and summarize sources. However, they are also notorious for not always using on factual information or rigorous research strategies. They are known for producing "hallucinations," a term used to describe false information created by AI systems to defend their statements. "Hallucinations" can be presented confidently and consist of partially or fully fabricated citations or facts.
AI tools have been used to intentionally produce false images or audiovisual recordings to spread misinformation and mislead the audience. Referred to as "deep fakes," these materials can be utilized to subvert democratic processes and are thus particularly dangerous. 
Additionally, the information provided by generative AI tools may not be current as some systems do not have access to the latest information. Rather, they may have been trained on past datasets and generate dated representations of current events and the related information landscape.

Bias

Another limitation of AI is the bias that can be embedded in the products it generates. These large language model systems are trained to predict the most likely sequence of words in response to a given prompt and will therefore reflect and perpetuate the biases inherent in the information they were trained on. An additional source of bias is the use of reinforcement learning with human feedback (RLHF) to refine generative AI tools. The human testers used to provide feedback to AI are themselves non-neutral. Accordingly, generative AI like ChatGPT is documented to have provided output that is socio-politically biased. It can also generate sexist, racist, or otherwise offensive information.       

Recommendations  

• Verify all of the information produced by generative AI. This includes checking the source of all citations the AI uses to support its claims to make sure the information generated by AI accurately reflects the contents of these resources.
• Critically evaluate all AI output for any biases that could skew the presented information. 
• Avoid asking the AI tools to produce a list of sources on a specific topic. These prompts may result in the tools fabricating false citations. 
• When available, consult the AI developers' notes to determine if the tool's information is up-to-date.
• Always remember that generative AI tools are not search engines. These tools simply use large amounts of data to generate responses constructed to "make sense" according to the algorithms they've been given and the data they've been trained on.

Learn More

• AI Was Asked to Create Images of Black African Docs Treating White Kids. How'd It Go? 
  Drahl, C. (2023, October 6). AI was asked to create images of Black African docs treating white kids. How'd it go?. NPR.
• Artificial Hallucinations in ChatGPT: Implications in Scientific Writing 
  Alkaissi, H., & McFarlane, S. I. (2023). Artificial hallucinations in ChatGPT: Implications in scientific writing. Curēus, 15(2), e35179–e35179. https://doi.org/10.7759/cureus.35179
• Fabrication and Errors in the Bibliographic Citations Generated by ChatGPT 
  Walters, W. H., & Wilder, E. I. (2023). Fabrication and errors in the bibliographic citations generated by ChatGPT. Scientific Reports, 13(1), 14045–14045. https://doi.org/10.1038/s41598-023-41032-5
• More Human than Human: Measuring ChatGPT Political Bias 
  Motoki, F., Pinho Neto, V., & Rodrigues, V. (2024). More human than human: measuring ChatGPT political bias. Public Choice, 198(1–2), 3–23. https://doi.org/10.1007/s11127-023-01097-2
• A New Chatbot Might Do Your Homework for You. But It's Still Not an A+ Student 
  Bowman, E. (2022, December 19). A new AI chatbot might do your homework for you. But it's still not an A+ student. NPR.
• Who Is Making Sure the A.I. Machines Aren’t Racist? 
  Metz, C. (2021, March 15). Who is making sure the A.I. machines aren’t racist? New York Times [Digital Edition].

Plagiarism

Plagiarism is typically defined as presenting someone else's work or ideas as one's own, including work generated by AI. Individual policies for using and crediting AI tools vary from class to class. Make sure you look at your syllabus and have a clear understanding of your professor's expectations. 

False Citations

AI tools can create false citations. Providing false citations in research, whether intentionally or unintentionally, violates academic honesty policies related to improper citation. AI tools such as ChatGPT have been known to generate false citations (hallucinations). Even if the citations represent actual papers, the cited content in ChatGPT or another tool might still be inaccurate.

Recommendations

• Know your professors' policies for AI use in every class. We cannot stress this enough.
• If AI tools are only permitted for topic development in the early stages of research, you might not need to cite them, but it's still important to check with your professor first.
• If you are providing commentary or analysis on the text generated by a chatbot and are either paraphrasing its results or quoting it directly, a citation is always required. You can find more information on citing AI tools on this guide.
• If you plan to publish your work, it is best to review that publisher's policies on the permitted use of Generative AI tools.
• Look up all citations and check to make sure they are accurate. If you're citing information from that source, to cite the original source rather than the AI-generated citation.

Learn More

**U.S. Copyright law related to the use of AI is still evolving. 

Do Generative AI tools violate U.S. Copyright Law?

There are currently several court cases directly relating to the unauthorized use of copyrighted material as training data for Generative AI tools. Individual authors, artists, and companies are suing OpenAI, GitHub, and other companies for using their work when training their AI products.
Some library content providers prohibit any amount of their content being used with AI tools. Do not download Library materials (i.e., articles, ebooks, infographics, psychographics, or other datasets) into AI unless you know our licensing allows this use.

Copyright vs plagiarism: Copyright violation is not the same as plagiarism. While plagiarism can be considered fraud if funding is involved, it is largely considered an issue of research integrity and ethics. There is currently no consensus over whether generative AI tools are engaging in plagiarism when they scrape data to generate content.
Can an AI tool retain copyright as an author? Copyright law currently has a human authorship requirement, and according to recent guidance, when an AI technology. What this means is that AI-generated art and text is not copyrightable on its own. 
Can I register my work with the U.S. Copyright Office if it was partially authored by an AI tool? This depends on the extent to which the AI tool is part of the creative process. The more human creativity involved, the more likely it is that you will be able to register your work with the U.S. Copyright Office. While you own the copyright to anything you create (or had a large part in the creation of), copyright registration is important as a public record of your copyright claim, which will be helpful to you if you are interested in licensing your work.

Learn More

Breaches of Privacy

There are multiple privacy concerns associated with the use of generative AI tools. The most prominent issues revolve around the possibility of a breach of personal/sensitive data. Most AI-powered language models, including ChatGPT, require users to input large amounts of data to be trained and generate new information products effectively. This translates into personal or sensitive user-submitted data becoming an integral part of the collection of material used to further train the AI without the explicit consent of the user. Moreover, certain generative AI policies permit AI developers to profit off of this personal/sensitive information by selling it to third parties.  

Recommendations

• Avoid sharing any personal or sensitive information via AI-powered tools. 
• Always review the privacy policy of the generative AI tools before utilizing them. Be cautious about policies that permit for the inputted data to be freely distributed to third-party vendors and/or other users. 

Learn More

• From ChatGPT to ThreatGPT: Impact of Generative AI in Cybersecurity and Privacy 
  Gupta, M., Akiri, C., Aryal, K., Parker, E., & Praharaj, L. (2023). From ChatGPT to ThreatGPT: Impact of generative AI in cybersecurity and privacy. IEEE Access, 11, 80218–80245. https://doi.org/10.1109/ACCESS.2023.3300381
• Why You Shouldn't Tell ChatGPT Your Secrets 
  Hunter, T. (2023, April 28). Why you shouldn’t tell ChatGPT your secrets. The Washington Post.
• Understanding the Ethics of Data Generative AI: Privacy Concerns and Solutions 
  Shafiq, M (2023, September 25). Understanding the ethics of data generative AI: Privacy concerns and solutions. LinkedIn.

If you use AI in your coursework, your instructors may ask you to describe how you integrated AI into your assignment. Use these reflective prompts to help you communicate about your use of AI.

• What tool(s) did you use?
• What prompts did you ask the AI tool to follow?
• How did you verify the information AI generated for you?
• How did you modify and use the AI output?

Learn More

AI outputs can be prone to misinformation and bias. Therefore, it is important to evaluate the accuracy of AI-generated content using lateral reading strategies. The strategies outlined here can help you to do so.

1. Carefully review the AI-generated content to identify specific facts, statistics, statements, and references that should be verified. 
2. Cross-check these pieces of information are true using reliable sources. If you cannot verify the information, the output is not reliable. 
3. Look for multiple sources that support the AI claims. Just one source supporting a piece of information is not enough. 
4. Evaluate the credibility of the sources you use to verify information generated by AI.
5. Verify citations provided by AI. Sometimes GenAI creates false citations. Search for the exact paper in an academic database. If you can't find the paper, it may not exist. Ask a librarian to help, as needed!

Lateral Reading & SIFT are useful techniques when evaluating AI generated content.

From MinnState's Educational Development and Technology department: