I'm worried about Chat GPT!

To be more specific I'm worried about my increasing reliance on Chat GPT (Artificial Intelligence) and how it's shifting the way I approach challenges in my work.

First off, I want to set the record straight. Since Chat GPT started gaining attention in November, I have been blown away by it. I have spent time both studying how to use it and also working on using multiple prompts to shape the conversation and get the best results from my interactions.

It started very binary in my interactions. "Question" -> "Answer", but has quickly evolved into really long and carefully crafted prompts, through a process of me understanding more and more about how to leverage this "Great Power". All this in just a couple of months.

I'm worried due to the following two factors:

• My increasing reliance on Chat GPT as being my work "copilot"
• My feeling of "cheating" in some of the things I am able to do with Chat GPT

Let's paint these two factors with examples:

Power Platform

I am a huge fan of citizen development and the rise of tooling and technology to enable this and have spent the last 18 months investing a lot of my personal and professional time in learning how to make the most out of this technology. In the last 6 months in particular I've used my new found skills to build production applications that are being used in real-world scenarios.

I saw a demo recently from a developer who is able to code, leveraging the Azure Open AI service to synthesise content and then generate a template from this. He talked me through the approach that he'd taken and I was convinced I needed to find a way of doing this for myself (but in a low code fashion).

After getting some help to setup the API call (due to some restrictions in the Power Platform environment) I was up and running and able to build prompts to generate responses from the GPT 3.5 turbo service.

There are two scenarios within this that I'm currently undertaking.

1. Using the input from the user, passing this against a scoring measure (with GPT3.5) and then returning an unbiased scoring across seven measures. This scoring is used to determine whether the input submitted is a high, medium or low impact initiative and to onboard those users onto a program of tailored content.
2. I am using input from a user, which is actually a reverse engineered prompt (ask below if you want to know more), to generate and fill in templates, that can be used to get first drafts stood up really quickly.

In both of these scenarios, the ideation, creation and actual implementation of the scenarios has been aided by Chat GPT.

To make a point here, "I would not have been able to achieve this prior to Chat GPT being a solution, without a Data Scientist + Developer.

A video project for a new platform

I have quite some experience form the past, particularly when I worked in Marketing with preparing videos. In this particular case, recently, I was asked to help out with a project where I didn't have a full understanding of the solution that the video would be produced around.

I hopped over to Chat GPT (with some inspiration from YouTube), and through a process of reviewing the slides that had been put together, shaped a prompt to request an outline script for the video. As per usual Chat GPT4 overachieved and gave me a strong foundation, through which I could present this structure to the project team and start gaining their input. For context, in the past I would likely have spent a number of hours, researching the topic, grappling with the context, and then trying to adapt all that to a script. In this case, I was up and running in less than 10 minutes! LESS THAN 10 MINUTES!

There's more. On Friday I was really tired, after a really heavy week. We had a last minute meeting, late in my evening around this topic. As two of the other team members we discussing elements of the script I was reworking parts of the script. I was using the live feedback of what folks in the virtual room were providing and using Chat GPT with this input, to adjust and shape new copy for the script of the video. I have to tell you now, it felt like cheating and I felt a little bit guilty for "working behind the scenes" in this fashion. I think this is the new reality, and predict this is going to be a process that many people will go through.

A customer presentation on Product Ownership

I was invited by one of the team working with a large global customer, a couple of weeks ago to talk about the role of the Product Owner. It had been some time since I had talked about this topic, in such a formal fashion. I immediately approached Chat GPT to help me to prepare the agenda and outline of the presentation.

Using this outline I had places to start leveraging existing content inside Microsoft, but I felt some of the key elements of product ownership, to explain in layman's terms were missing. I had an idea of using an analogy to paint these examples and shared this with Chat GPT. It was then able to go on and create example to quickly and easily align to these. Combining this together with Midjourney, I was able to make a presentation in 4 hours (from scratch) to give a 1 hour presentation on this topic. Again this is about 1/4 of the time I would normally spend.

Why so worried?

You might have read the above and think "well that's a great use of technology", but there's more here.

What's surfacing in my mind, is that the content, my approach, the outputs, for the work that I do is changing, and I am quickly adjusting the way I work to adopt this new approach. I'm feeling guilty for being bale to work in this way, and that reason for feeling guilt, I believe is being motivated by the primary factor of not seeing others yet leveraging the AI in the same way.

I worry, that the work I was doing just 6 months ago is becoming unnecessary and that we need to move past talking about AI and it's potential impact to actually working together to apply AI in our work, and identifying where the expectations would be between "using AI" and "using your own brain".

I think this expectation setting will help folks to start to anchor their learning journey against something well thought through and defined. There is significant disruption coming, and it shouldn't simply be "survival of the fittest". We need to work collaboratively to support each other and our learning journey and continue to invest in identifying opportunities for others to leverage AI. I commit to investing time in teaching and supporting those I interact with to leverage AI to support them in their work, so I can get rid of the feeling of cheating! ??

Thanks for reading!

Notes:

• I don't use company data or PII in my interactions with Chat GPT. I have found a way to paint my picture of my needs without this.
• I love AI imagery. I use it often as I feel it helps capture the impact of the words I want to use and is great for analogies.
• If you have questions or would like me to explain anything additional please let me know.