Artificial Intelligence is the capability of computer systems or algorithms to imitate intelligent human behavior. Generative AI is artificial intelligence that is capable of generating new content (such as images or text) in response to a prompt by referencing a large database of text or other information.
Generative AI are trained using large amounts of data. By applying algorithms that allow them to identify patterns in the data, they can generate their own content based on those patterns. Generative AI that generate text are also called large language models (LLM).
Some of the popular Generative AI tools currently available include ChatGPT, Microsoft Copilot, Gemini, DALL-E, and Midjourney. Many apps now include some kind of AI features as well.
Because humans create and select the data and algorithms the generative AI is trained on, there are limits to what data the AI has available. Output can thus reflect biases in the data, or even create nonsensical or inaccurate responses that are called hallucinations.
You may be familiar with autofill/autocorrect tools, such as Microsoft Word's spell checker, Google Search boxes that finish your sentences, or predictive text when messaging friends.
These are all types of artificial intelligence tools, however they are NOT generative! These tools do not create brand new or original material based on a prompt. When your professor tells you not to use generative AI, it is generally OK to still use spellcheck and autocorrect, but you should always check with them to be sure. Keep in mind that Grammarly, which is a spelling and grammar checker, also has a generative AI product called Grammarly Go.
Additionally, language translation tools like Google Translate are not considered generative as they are not creating new material, they are simply translating existing material.
AI can be useful for many different things, like brainstorming, topic creation, idea generation, or suggesting keywords to use in library database searches. It also has specific applications within many majors and fields.
Each course you take will have its own policy on AI based on that course's learning objectives. ALWAYS check with your professor before using AI in any assignment! Remember, that submitting anything written using generative AI, especially without the explicit permission of the professor, could put you at risk for accusations of plagiarism.
Generative AI has limits to its available data, and its output can reflect cultural biases. When an AI sources information from across the internet (or even, sometimes, from a proprietary database), the result can be output reflecting racism, sexism, homophobia, or outdated or false information.
Here is a simple example of ChatGPT being unable to answer a question properly:
Hallucinations happen when the AI makes up its own answer that sounds persuasive, but is actually nonsensical or incorrect. Furthermore, AI can manufacture false article citations for studies that don't exist, so be sure to follow up thoroughly on studies and sources.
It's always a good idea to evaluate generative AI output. Use techniques such as Lateral Reading to evaluate generated output by crosschecking with outside sources.
Whenever you use AI output, you must cite it just as you would any other source. Citing sources is a crucial part of the research process, and it can be tricky figuring out how to cite a non-traditional source like AI and ChatGPT, especially since those outputs are not retrievable by other users. Fortunately, many citation styles now have a way to cite AI.
Here are guides to four styles used at WNE:
In addition to the citation, some professors may want you to include a copy of the prompt you used and a transcript of your conversation with the AI.
Generative AI can play a role in the scholarly process. Tools like Research Rabbit, Scite and Elicit can assist with discovery, literature mapping, and evaluating sources. Professional organizations such as the APA and COPE have developed guidelines for using and disclosing AI within published content.
Generative AI also has uses within the classroom. The Center for Transformative Teaching at University of Nebraska-Lincoln has an excellent guide on using AI in teaching, including example course policies, prompts, and assignments.
“ChatGPT Has Changed Teaching. Our Readers Tell Us How.” The Chronicle of Higher Education, 11 Dec. 2023, https://www.chronicle.com/article/chatgpt-has-changed-teaching-our-readers-told-us-how.
David, L. (2019). Understanding artificial intelligence ethics and safety. The Alan Turing Institute. https://www.turing.ac.uk/sites/default/files/2019-08/understanding_artificial_intelligence_ethics_and_safety.pdf
Dee, Catherine. "Large language models (LLMs) vs generative AI: what’s the difference?" Algolia. https://www.algolia.com/blog/ai/large-language-models-llms-vs-generative-ai-whats-the-difference/
Guidance for Generative AI in Education and Research - UNESCO Digital Library. https://unesdoc.unesco.org/ark:/48223/pf0000386693. Accessed 29 May 2024.
Hartman-Caverly, Sarah. Library Guides: Hidden Layer: Intellectual Privacy and Generative AI: Hidden Layer Workshop. https://guides.libraries.psu.edu/berks/AI/workshop. Accessed 29 May 2024.
“How AI Reduces the World to Stereotypes.” Rest of World, 10 Oct. 2023, https://restofworld.org/2023/ai-image-stereotypes/.
IBM Technology. What Are Generative AI Models? 2023, https://www.youtube.com/watch?v=hfIUstzHs9A.
Key AI Terminology | GSA - IT Modernization Centers of Excellence. https://coe.gsa.gov/coe/ai-guide-for-government/what-is-ai-key-terminology/. Accessed 30 May 2024.
Maheshwari, Pradip. “How to Use Scite AI: A Smart Guide for Researchers.” Open AI Master, 31 Aug. 2023, https://openaimaster.com/how-to-use-scite-ai/.
Subbaraman, Nidhi. “Exclusive | Flood of Fake Science Forces Multiple Journal Closures.” WSJ, 14 May 2024, https://www.wsj.com/science/academic-studies-research-paper-mills-journals-publishing-f5a3d4bc.
What Is AI Model Training and Why Is It Important? https://www.oracle.com/artificial-intelligence/ai-model-training/. Accessed 30 May 2024.