"

Guided Critical Thinking

Marcus Lacher

Bias-Aware Prompting

Bias-aware prompting involves crafting AI inputs that aim to minimize internal and external biases in the responses generated. By being intentional with language and structure, educators and students can guide AI tools to produce more balanced, inclusive, and credible content. This practice supports ethical scholarship and helps develop critical thinking around how information is created and presented.

Example: Business Technology Trends Research Project

Goal

This is an ethical example of using generative Artificial Intelligence to support the creation of OERs. The main goal of this endeavor is to help students engineer AI prompts that generate responses that reduce or eliminate internal and external biases, while also identifying relevant and reliable sources of professional research. These topics can be controversial for potentially political, economic, and social reasons, but the students need to find credible sources that support their positions and acknowledge the counterarguments of the topic. It is important to remember that it is not essential to create the perfect AI prompt, but to consistently interact with GenAI prompt to get the desired output. The more you experiment, the better off you are.

How It Was Done

Instead of building a complex chatbot, this project began with a simpler, strategic approach. A set of carefully designed GenAI prompts was created to help students filter out biased, outdated, or unreliable information. These prompts guide students in generating quality content from credible sources, which they can then evaluate using their own critical thinking to support their research goals.

  • Traditional, unguided internet research can produce data overload. Much of the data generated does not qualify as information. It’s just data. Using a GenAI prompting template or AI Assistant can guide researchers to collect quality information from verifiable sources that can easily be scrutinized for potential bias and relevancy.
  • Generative AI is helpful because it encourages refinement of research by suggesting additional prompts, whether they are user-defined intervention or the chatbot recommended prompts.

Results

Copilot guides the user rather than simply providing answers the user can copy. My human intervention steered Copilot towards finding specific examples of ethical issues with the research topic. This provided authentic, recent examples that offered convincing evidence to support a research hypothesis.

 

Considerations

Potential issues

  • All generative AI tools, including ChatGPT and Microsoft Copilot, are vulnerable to data security, privacy issues, and various ethical concerns. Both of these tools can suffer from issues connected to AI bias and may occasionally surface incorrect responses to questions.

Best Tools for the Job

  • While any leading AI Chatbot can provide similar results, I prefer Copilot. Why?
    • Since Microsoft’s Bing search engine powers Copilot, the tool has access to the internet. Therefore, Copilot can answer questions with more accurate and up-to-date information.
      • When researching trending technology topics, research to real-time internet information is crucial!
    • The free version of ChatGPT is not connected to the internet and, consequently, only has access to the information it was trained on, limiting its knowledge base to up to January 2022.
    • Copilot includes footnotes with every response that cites the website the chatbot pulled its answer from. Cited sources make it easier to further your research and confirm the accuracy of the chatbot’s response.
      • ChatGPT will provide sources IF the user remembers to prompt ChatGPT.
    • OpenAI trains its LLMs on data from customer conversations with ChatGPT unless you directly opt out of this process or use the Team or Enterprise version of the tool. Microsoft, on the other hand, says that it doesn’t use customer data to train Copilot tools at all. Prompts, responses, and data accessed through Microsoft Services aren’t used to build on the functionality of Copilot in any tool.[1]

  1. https://www.uctoday.com/unified-communications/microsoft-copilot-vs-chatgpt-which-is-best-for-you/

About the author

License

Icon for the Creative Commons Attribution-NonCommercial 4.0 International License

Guided Critical Thinking Copyright © 2025 by Marcus Lacher is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License, except where otherwise noted.

Share This Book