Generative AI tools (Gemini, CoPilot, ChatGPT and more)

Google Gemini, Microsoft CoPilot, and ChatGPT are examples of "generative artificial intelligence" or "large language models."  These tools differ from a Google search - where you are given results to a specific search inquiry. Instead GenAI tools create “new content” by predicting the word most likely to come next (e.g. based on publicly available Internet sites including Wikipedia and YouTube transcripts).

Large Language Learning Models (LLMs), like Google Gemini, ChatGPT and Microsoft CoPilot, are designed to model human language. They use mathematical models to predict what the next word is most likely to be based on what you are asking for, but because they use natural language, it "feels" like they are thinking. GenAI tools are trained on HUGE dataset of publicly available Internet sites including places like Wikipedia and Youtube transcripts. Evaluation of the quality of the sources is not a part of the GenAI output. 

One main way users interact with these GenAI tools is to ask it a question, or give it a prompt and receive a quick answer.

Using GenAI tools as a student

• First thing to know, is each class or project might have different rules or expectations on AI use. Each instructor gets to decide how they are using or not using these tools in their classes. Check each syllabus for your instructor's guidelines. If you are unsure, it is your responsibility to ask the instructor or professor about it. Here are some examples of how it might be in the syllabus. 
• When using AI tools, a good guideline is to keep track of how you are using the tools then consider drafting an AI statement or acknowledgement - this simply means describing how you have used Generative AI in the process of creating your paper or assignment. If you are using an AI-generated text, images, or other media, be sure to cite it as a "source" so your readers/instructors know the origins of the material. 
• Graduate students, be sure you have read Guidance on Generative AI in Graduate Education.
• Lastly, use the highest ethical standards when using AI:
  • Originality: Ensure all work submitted is YOUR own, with your unique voice and ideas. When permitted, you can use AI tools for help but not to do the work for you, unless your instructor says it’s okay to use AI in a specific way.
  • Acknowledgment and citing: Follow your instructor's rules for citing or acknowledging AI use in your course. If you’re uncertain, ask your instructor about it. APA, MLA, and other citation styles now have guidelines and examples for citing generative AI work. Ask a librarian for help! 
  • Share with caution: Any information you share with an AI tool is no longer private and may be used in ways you can’t control. Using UMN licensed tools (e.g. you login with your UMN email) like Google Gemini and Microsoft Copilot do include data protections but never share sensitive, personal or confidential information about yourself, others or your organization with AI systems.

Using GenAI tools at the UMN

Publishers typically prohibit (through licensing) upload of research, magazine, and newspaper articles (including those used for course readings) to third-party services like AI tools. Do not upload materials produced by others or licensed materials unless you have explicit permission or the material is covered under a known exemption or license. For more information read the Libraries Acceptable Use of Electronic Resources.

UMN Policies

Ethical concerns of using GenAI or other AI tools:

There are a variety of valid and serious ethical concerns with this technology broadly including:

• exploitative labor practices in the creation of and training of tools
• social, cultural, economic impacts (many of which are unknown yet) 
• Bias and discrimination (bias amplifier) in the content the tools are trained with
• Privacy and copyright in terms of “sharing” material with a tool with extreme caution. Also ethical considerations of the copyright owners of the content the tools are trained with
• Environmental impact (energy, water, etc.) of creating, training and using tools 

It is useful to be aware of these ethical challenges so you can be critical and make decisions on using the tools.

Challenges of using GenAI or other LLMs 

• There is a mix of correct and incorrect information
• Has limited knowledge of very current events in the world (e.g. last week, last month, etc.) 
• Likelihood of biased content is high - especially for controversial topics (these tools have been called bias amplifiers) 
• Privacy concerns -- what is the company doing with the data it collects from users?
• Beware of asking for any information that would have big consequences if it was incorrect (such as health, financial, legal advice, and so on). It has a tendency to make up answers or give a mix of correct and incorrect information, but still sound very confident.

Benefits of using GenAI or other LLMs 

• Can provide simple explanations to well known, non-controversial topics. These tools can help to explain information in ways that are easy to understand.
• Can help to get unstuck or move on to the next steps of a task
• Can create a list of keywords, search terms
• Helping write or debug computing code
• Summarizing and outlining texts - but remember the tool doesn't know the most important point but instead is guessing the next best word - so verify that the summary is accurate. 

What are prompts? Prompt engineering? 

Prompts are the things you write into the tool to try to get it to do what you want. Better prompts can help you to try to get better outputs. These tools need very specific instructions, and they need you to verify/critically evaluate the information or output they give you and give them feedback so they can get better at the task. Learn more about prompts or a course like Prompt Engineering for Chat GPT.