An Interview with ChatGPT on Asking the Right Questions

ChatGPT gives wrong answers—for tips on preventing this, we turned to the source

February 24, 2023 | Denise Gifford

ChatGPTThe advances in artificial intelligence as demonstrated by OpenAI’s new ChatGPT chatbot are truly amazing. It suddenly seems like anyone can get answers about any topic, or can use the chatbot to write essays, compose poems, solicit advice, and create all kinds of content at speeds unattainable by humans.

While most of us are familiar with chatbots, it is the generative artificial intelligence (GPT stands for generative pretrained transformer) that marks a huge advance. The model has been trained to interact with users in a conversational way, and that has encouraged people to try it out and see what it can (and can’t) do. Within the first five days of ChatGPT’s launch in late 2022, over 1 million people signed up to use it. Today it has 100 million active monthly users. ChatGPT is the newest way to get answers, like Google before it. But, just as with search engines, there are many kinks and quirks to the chatbot’s functionality. The usefulness of its results depends upon how we ask our questions.

Asking the right question can be a challenge, whether we’re using technology or conversing with another person. The art and science of question construction has been studied by researchers for years. It is a key skill in journalism, law enforcement, teaching, management, and just about any pursuit in life. The questions we ask are so important. The author Warren Berger, who has written extensively about the power of inquiry, notes that questioning is essential for critical thinking, especially since “much of the information coming at us today is unverified, unreliable, and quite possibly untrue.”

So what does ChatGPT itself have to say about asking the right questions?

I asked it:

How can I ask better questions on ChatGPT?

Chat GPT responded:

When asking questions on ChatGPT, you can improve the quality of your questions by following these tips:

  1. Be specific: Instead of asking a general question, try to be as specific as possible. This will help you get more accurate and relevant answers.
  2. Use proper grammar and punctuation: Make sure your question is well-formed and easy to understand. This will help ChatGPT understand your question better.
  3. Provide context: If your question is related to a specific topic or situation, provide some context to help ChatGPT understand the context of your question.
  4. Be clear: Make sure your question is clear and easy to understand. Avoid using jargon or technical terms that others may not be familiar with.
  5. Ask open-ended questions: Instead of asking yes or no questions, ask open-ended questions that allow for more detailed answers.
  6. Follow up: If you don’t understand the answer to your question, don’t be afraid to ask follow-up questions to clarify.
  7. Be concise: Avoid asking multiple questions in one go, try to keep your question concise and focused on one topic.
  8. Be polite: Always be polite when asking a question. This will help you get more helpful answers.

These recommendations generated by ChatGPT turned out to be a pretty standard list of “how-tos” for asking better questions. But, of course, it was simply doing what it was designed to do. It’s a neural network that learns from the vast amount of data on the internet.

Sometimes its answers don’t make sense, and sometimes it is just plain wrong. For example, ChatGPT incorrectly identified the second largest country in Central America (other than Mexico) as Guatemala, rather than Honduras. And, looking at the list it gave me, ChatGPT recommended that I be polite when asking it questions—nonsensical advice for dealing with a chatbot.

It even gave me this disclaimer:

Remember that ChatGPT is a language model, it can generate a wide range of answers based on its training data, but it’s not always accurate or complete, so use your own judgement when evaluating the answers you get.

This new, amazing technology can provide plenty of answers, but it’s still up to us not only to get the initial questions right, but to judge the validity of the answers. In other words, it’s important that we continue to question every step of the way, including assessing the answers. We’re responsible for determining the quality of the response. As the disclaimer notes: “use your own judgement.”

The AI technology that can sweep through data, recognize patterns, and learn from what’s available to it online is still learning, after all, from the internet. Think about that when you use it.

  • About The Author
  • Denise Gifford is the managing editor of She co-founded InfoWorks International, a consulting firm that trained thousands worldwide in project management, leadership, and related business skills. Prior to heading InfoWorks, Denise worked in sales and marketing management and as a consultant to the financial services industry. She holds her MBA from Northwestern University’s Kellogg School of Management.