Critical Evaluation of AI
One of the best ways to evaluate information that you get from an AI tool is to build lateral reading into your process. Lateral Reading is when you verify the claims of one source by comparing it to other sources on the topic. Every piece of writing includes some of the preconceived notions of the author. As the reader, part of your job is to identify varying perspectives and think about how they overlap. It may sound easy, but because we all come with cognitive biases, it's actually pretty difficult to do this. AI can make this especially hard because it writes with authority and uses human-like speech patterns.
Here are some strategies to help you use lateral reading in your own work:
- Choose an AI tool that includes sources with its answers. Perplexity is a good choice. After you ask a question, you can click on the number bubble to read the original source.
- Search the claims in an academic search engine or database like Google Scholar or DELCAT and then look for similarities and differences.
I know that lateral reading sounds like a lot of work when AI was supposed to make your research easier. To simplify this process, try taking notes on important terminology from your AI summary, this process is called background research. Then, you can use those terms to search for reliable information in an academic database, this process is called foreground research. You can also combat misleading information by intentionally prompting your AI tool to include multiple perspectives. You can ask your AI tool something like this:
- "What are the competing viewpoints on [insert your topic]?"
- "Why might someone disagree with [insert your argument or point of view]?
In the next portion of this tutorial, we'll explore the importance of your ideas in writing AI prompts.
