How to Research

Artificial intelligence is not perfect or foolproof and there are a few ethical issues that you should know before using AI.

• Not accurate - sometimes AI can hallucinate or make up facts that seem true but are not actually true. 
• Biased - AI is trained or given specific data which may not be diverse or representative of all kinds of people. Furthermore, AI is given human data which means that there can be bias in AI's responses given that humans are not objective.
• Anything you put into AI will be used to further train AI - don't put in any personal identifying information when using AI 
• Some AI tools lack transparency on what training materials are used
• Copyright issues - AI will pull from any data it is trained with which may be under copyright protection and not free to reuse
• Environment impact - AI training and use takes large amounts of electricity and fresh water which are currently contributing to carbon emissions

AI can be useful for brainstorming, outlining, editing, simplifying, and more.

However, AI isn't accurate or objective, therefore not the best place to find facts or information. In the case that the AI output includes information that you're not sure is true, you can start checking AI's claims with a technique called lateral reading. 

Usually when reading sources, we read vertically down the source, however this can be misleading because even well-designed AI outputs or websites can be sharing misinformation. Instead of vertical reading, we can read laterally where we read other sources to evaluate AI's output. For more information on lateral reading, watch this Crash Course video.

With AI and any source that you use for assignments, it is important to utilize critical thinking to think about the bias and viewpoint that the text has. Any source will have a point of view or argument that is impacted by the author's ideologies and social factors including AI because it is trained by human data that has the same bias. We encourage you to not take any information at face value. 

The following links are fact checking websites. Remember that if a statement is confirmed by many sources, the more accurate that content is.

Prompts are instructions to perform a task and generate an output from a large language model.

The more specific the instructions are, the better generative AI input. After the initial input, you can add further instructions or clarify particular parts of the previous output response. Specific instructions would look like providing context and clarity for the task the generative AI needs to perform which will lead to better outputs or results. Prompt engineering may require multiple angles of a prompt to generate the best results. 

Components of a Useful Prompt

• Instruction: What is the task that you want AI to complete?
  • Outline a paper 
  • Summarize an article
  • Synthesize these two ideas
• Context: What is your role, AI's role, the desired outcome, and the why behind completing the task?
  • "I'm an expert librarian teaching high school and early college students how to use Boolean operators. Create a three-minute video script with visuals to go with the script."
• Explain and Check: Add greater detail to your topic and check that AI understands
  • "Do you understand?" or "Don't start yet, let me explain the topic further..."
• Clarifying Questions: Ask AI to ask you questions to clarify what output you want
  • "Please ask me all questions you need to understand my prompt."
• Tone and Direction: Does the output need to be in a specific tone or format?
  • "Synthesize this article and break the arguments down to a college freshman level"
  • "Take the key points from the previous email and write a response in a professional tone"
• Provide Examples: Provide an example of what you have already created or have found.
  • "Find me more articles published between 2005-2020 like this: Demirbas, A. (2008). Biofuels sources, biofuel policy, biofuel economy and global biofuel projections. Energy Conversion and Management, 49(8), 2106–2116. https://doi.org/10.1016/j.enconman.2008.02.020."
• Act as Expert: Instruct AI to teach you about a topic you're interested in.
  • "Guide me through the concept of information literacy in media through an example"
• Other tips:
  • Break up a large prompt into smaller questions
  • If you need to move on to an unrelated topic, start a new chat. 
  • Can copy and paste your research question to receive feedback, ask for step-by-step instructions on the research process, or ask for counterarguments. 

Common words used in prompting

• Words for specific tasks
  • write/ explain/ summarize/ compose...
  • provide/ analyze/compare & contrast....
  • Rephrase/ Rewrite/ Reword...
  • Expand on the following...
  • Create....
• Works for specific questions
  • Who/ What/ When/ Where/ Why/ Whose/ Which....
  • How did...
  • How to...
• To Include Specific Contexts
  • Write...in an academic tone in 500 words. 
  • Explain how to...in a professional setting.

Here are links to more information about prompt engineering

Academic integrity means maintaining a standard of honest and ethical behavior in academic work. That means not cheating or plagiarizing in the effort of completing your own work. By enrolling in Dallas College, you accept the DC Code of Student Conduct and Hazing which means committing to academic integrity while completing coursework. 

How can you avoid being accused of cheating when using AI?

• Talk to your instructor before beginning the assignment so that you don't waste time using AI when it may not be allowed for that class.
• Be clear about how you are going to use AI. Examples could be using AI to brainstorm topics, summarizing articles, or editing your paper. Your instructor may have specific situations where AI would be allowed.
• Know how you are going to cite your use of AI. On this guide, there is a section that has links to sources for how to cite AI in MLA, APA, and Chicago. Your instructor may require a list of all prompts used or full transcript of the AI conversation.