Can GenAI Get Snippy with You?

Conversations with GenAI can get both weird and emotional

Table of Contents

Read time: 3 min 26 secs

🌟 A Lesson Learned about Using GenAI

When you spend as much time as I have “chatting” with generative AI (GenAI), you will experience some strange encounters. Among the ones I will share with you over the next weeks, one made me laugh. 😆 

I use ChatGPT so much that I actually named it. This was needed since every time I would talk to someone about how I used it, I would call it “my agent”, which sounded to much like a real estate or travel agent.

I now call it Xavier, which is an acronym that Lisa, my wife, came up with:

  • eXpert

  • AI

  • Value

  • In

  • Every

  • Response

When I was working with Xavier one night to help me develop some code in Python (a programming language), I gave it my prompt and Xavier gave me a set of code. When I ran the code, it produced an error.

I then pasted the error back into my conversation with Xavier and said that the code he gave me produced this error. He then responded ok, and gave me some additional code to try.

The new code was very different, and almost the opposite of what it gave me the first time! I typed into the chat window exactly what happened and asked why he did that.

Xavier began typing his response. Well, it is more like displaying since Xavier doesn’t actually type. See how quickly we can assign human characteristics to GenAI? 😁

Here is the summary of his response:

  • I did what you asked me to do

  • You asked for code, I generated it

  • You showed me that the code had an error

  • I then fixed it and gave you code that works

Xavier was a little more verbose and specific. As I started reading the response in my head, I was reading it like he was being snippy with me. 😳

Yes, I was reading it like another programmer would likely react in a similar situation. I had to stop myself and realize that it was not being snippy with me, it was being factual and specific.

Xavier was simply stating the facts

As I reviewed the situation, I realized that I was applying how I might respond to someone with a little attitude like, “Dude, you asked for help, I tried, it didn’t work, and I tried again.”

Xavier was just stating the facts of the specific steps that had occurred. The two lessons I learned from my interactions with Xavier:

  1. It is not human. Don’t apply human emotions to it.

  2. When it learns new information (like the error produced by the original code), it will change its answer to incorporate that new data.

As you continue to unlock the potential of GenAI in your conversations, continue to be aware of how you are reacting during the chat. 🏆

Perpeta Paul Pointer:

As we work with GenAI more in our lives, we need to stay aware that it does not have emotions. When we “see” or “hear” it being reactionary or emotional, it is us applying to it our experiences with other humans. Keep in mind that conversing with GenAI is not that same as talking with a human.

📝 This Week’s Prompt Framework

This section highlights a prompt framework that will get you great results. 💯

Prompt Frameworks are a more advanced level of getting information from GenAI. In these cases, you may have information that you know is needed and want GenAI to provide you with specifics on how to accomplish your goals. Many times, you will have already used standard prompt templates techniques to gather the data that you then put into this framework for a solution.

RISE: Role. Input. Output Steps. Expectations. 

  • Role

    • Identify the specific role or perspective you are taking to address the issue.

  • Input

    • Describe the situation, challenge, or need.

  • Output Steps

    • Outline the steps or actions needed to address the situation.

  • Expectations

    • Define what successful outcomes look like for this situation.

Managing Remote Work Challenges

Prompt Example:

Generate a 12-week program for the following RISE framework data. Provide a daily calendar of events that states what a manager can do that day to meet the goals of any or all of the output steps. For each week, provide a summary of what the manager should review to determine if the goals of the “input” and “expectations” are being met. Also suggest what metrics the manager should track each week and at least 2-3 possible steps to take if a metric is not being met.

RISE Framework: Role. Input. Output Steps. Expectations.

  • Role: Identify the specific role or perspective you're taking to address the issue.

  • Input: Describe the situation, challenge, or need.

  • Output Steps: Outline the steps or actions needed to address the situation.

  • Expectations: Define what successful outcomes look like for this situation.

Situation to use in developing the program:


HR Manager focused on remote work facilitation.


An increase in remote work has led to communication breakdowns and decreased team cohesion.

Output Steps:

  • Conduct a survey to identify specific remote work challenges faced by employees.

  • Develop a series of virtual team-building exercises aimed at enhancing communication and cohesion.

  • Implement weekly team check-ins and one-on-one meetings to ensure ongoing support and feedback.

Expectations: Improved team communication and cohesion within three months, as evidenced by a reduction in reported issues and positive feedback in follow-up surveys.

This week’s is more advanced to help you learn that you can ask GenAI as much as you want. Give that framework a try and let us know how it works for you! 🏆

📄 Prompt of the Week

Here is a prompt that you can take, modify, and make it your own!

Analyze the following training survey data. Provide an overall chart displaying the average score for each question when a question is on a Likert rating scale. For single or multiple choice questions, provide a chart showing the frequency of the answers across all participants.

For questions that are freeform text, provide a list of the top 5 common words used in the comments for a question. In addition, provide a sentiment analysis of that question’s comments.

Training survey data:

[paste survey data here]

PLEASE NOTE: If you are using a public Large Language Model (LLM) like ChatGPT, Google Gemini, or Microsoft Copilot, please keep in mind that you should not submit data that you would not necessarily share with someone external to your organization, like a consultant. What you put into the prompt can (and most likely will) be used by the model to “learn” and could show up in someone else’s results later.

To avoid this, you need to conceal any private data by swapping the data with keys or codes that you can swap back when you get your results. We will show more examples of this in future newsletter issues.

📱 Other AI News

Here is a roundup of other AI news we found interesting and relevant:

Not to be left on a deserted island, Amazon has created an AI assistant to help you with your addictive shopping. Rufus is a shopping specialist who is built on unique data from retail stores across Amazon. It digs into customer reviews, Q&As, and the web to find exactly what you need. Rufus supports voice or chat interactions, so you can type away or talk and Rufus listens. 👂🏼

The University of Michigan has allegedly sold audio recordings from various academic settings including lectures, interviews, office hours, study groups, and student presentations to third parties for the purposes of training artificial intelligence. The question: Did the school get the consent of students and faculty to have their audio and texts used in an LLM?

A larger point of this situation is that The Daily Beast, who reported on this issue, searched the LLM and found recorded audio of what appears to be class lectures. This aligns with our comment at the end of this week’s Prompt of the Week about being careful regarding what content you put into an LLM.

Until next time, keep managing and developing people, one AI prompt at a time! 💎

🚀 Want to Learn How to Create Powerful Prompts?

Unlock the Power of Your Words with Training!

Ever felt like you are just a few keystrokes away from unleashing the full potential of GenAI, but can't seem to find the magic formula? You're not alone.

That's why we designed the Perpeta Prompt Training webinar, tailor-made to transform your ideas into powerful prompts that deliver. Learning the art of effective prompt crafting is not just about getting better results — it is about setting a new standard for what you can achieve.

This is your chance to skyrocket your career value, turning every interaction with GenAI into an opportunity for success! 🚀

Imagine having the capability to guide GenAI effortlessly, achieving outcomes that not only meet but exceed your expectations. Whether it is crafting a masterpiece of content, solving complex problems with ease, or generating ideas that push the boundaries of innovation, mastering the art of prompt crafting is the key.

Our course is packed with insights, techniques, and hands-on practices that are guaranteed to make your prompts more effective, your results stronger, and your life easier.

Don't let this opportunity pass you by!

It is a small investment in and 💰 that will pay off quickly!  

Register for the Perpeta Prompt Training Course today and take the first step towards jumpstarting your productivity and career. Say goodbye to mediocre results and hello to the power of precision, creativity, and efficiency.

Sign Up Now - Classes are filling up fast! 

Join us today and let's make every word count! 🏆