IT’S NOT TOO LATE TO LEARN ABOUT ARTIFICIAL INTELLIGENCE

Conversations about AI and generative AI have taken over most podcasts, blogs, and workplace Zoom conversations. At this point, you’re either using AI, or you’ve waited to experiment because you were unsure, and now it seems like everyone is light years ahead of you.

They’re not.

Well, I’m sure someone is.

But it’s not too late to learn about generative AI and how you can use it to amplify the work you’re already doing.

WHY LEARN ABOUT GENERATIVE AI?

You’ve been incorporating AI into your life for years, perhaps decades even.

Technology has always been evolving and with it, how the people who use it accomplish their daily work.

I don’t know if people were skeptical of tractors that replaced horses, the printing press, electric mixers, accounting software, the internet, or even the Google search engine – but I do know that each piece of technology has changed the way people live and work.

It’s time to get your feet wet in the generative AI waters.

The way it will impact your daily life may not be clear yet, but I have no doubt that the time is coming.

So, let’s start by splashing around in the shallow end, shall we?

WHAT IS GENERATIVE AI?

In a YouTube video of her Turing Lectures, Mirella Lapata from the University of Edinburgh describes Generative AI as a type of artificial intelligence that involves creating new data or content.

The computer program does not create this original content without a cause. A user must prompt the artificial intelligence to take action.

That’s where you come in.

WHAT ARE AI PROMPTS?

An AI prompt is the action the user tells the AI tool to take. For example, a user may prompt ChatGPT to “write a cover letter to accompany a resume for a marketing assistant.”

ChatGPT will then create original content based on that prompt.

Prompt writing in and of itself is becoming an art.

In the example I gave above, you may have questioned how ChatGPT would provide a personal letter.

It wouldn’t.

David Gewirtz says that sharing background information to set the proper context is one of the 5 steps to writing better prompts in this ZDNET article.

The prompt should include background on the applicant, details about the hiring company, data on letter format and writing style, etc.

AN EXPERIMENT

I’ve conducted a mini experiment to discover how different AI tools (chatbots) respond to prompts and generate content.

I have provided three chatbots with the same prompts. We’ll review their similarities, differences, accuracy, and the amount of prompt engineering needed to produce a suitable piece of content.

All three tools I am testing have free versions you can experiment with:

Gemini – Google

ChatGPT – OpenAI

FlowGPT – a fun character-based generative AI tool

PROMPT 1: SOCIAL MEDIA POST

Marketers and small business owners are using generative AI to help write social media captions, newsletter copy, and ads.

The first prompt I used asked our three chatbots to act like a Social Media Manager and write a social media post that targets millennial women in the workplace and teaches them how setting boundaries can improve their life.

Gemini

ChatGPT 4

FlowGPT – Kim Kardashian

THE RESULT

The most interesting finding to me is that the AI tool meant for fun over professionalism, FlowGPT, provided the best outcome after one prompt. It was short and written in a good style for Instagram. Although, I can always do without the ‘boss babes’ lines.

Gemini and ChatGPT provided captions that were entirely too long on the first try.

After asking them to cut the number of characters they provided in half, ChatGPT provided a better result. Gemini shortened their post down to the bare minimum.

All three tools provided accurate descriptions of how millennial women can improve their lives by setting boundaries. In their own ways, they each recommended that women learn how and when to say no to create healthy boundaries.

I wouldn’t use any of these captions as-is, but they are a great foundation for a social media reel or carousel idea. I also find that AI tools go very heavy on emojis, which I’m not a fan of.

Remember, you’ll want to give the chatbot information about your voice, company, and your specific audience when you ask it to help you write.

Prompt 2: Weight Loss Plan

Generative AI and LLMs (long language models) can do more than help you at work. People are using them in their everyday lives.

You can experiment with AI in your kitchen as you provide ingredients on hand and ask for possible recipes, use it to build itineraries as you travel with the family, and utilize it to improve your health.

For the second prompt, I asked all three chatbots to create a workout routine that would help me lose 12 pounds in 2 months without the use of equipment.

This time around, you’ll notice I received very different responses.

Gemini

ChatGPT 4

FlowGPT – Kim Kardashian

The Result

So, we need to address FlowGPT.

Ms. Kardashian did not want to respond to this question.

The chatbot did say it needed more information, but after receiving the information, it still did not want to contribute to the betterment of my health.

Ironic coming from a Kardashian? 🤔

Gemini and ChatGPT did their best in the first round but I knew they would need some more data. I provided height, weight, and age to receive a more detailed response.

Gemini’s response was short. Gemini provided one daily workout that would be used for the entirety, except rest days. This AI chatbot made adjustments after receiving the extra data.

However, Gemini responded with adjustment notes to its original plan, not a new detailed routine.

ChatGPT was far more detailed in its response.

ChatGPT gave me a weekly overview, a workout for each day of the week, and nutrition tips!

After I added the information ChatGPT would need to calculate BMI and other potential weight loss variants, the tool gave me a fully updated plan.

Based on the details I provided, the plan included workouts appropriate for the participant’s potential fitness level. ChatGPT also assumed this plan was now for a person on a journey back to fitness.

Rude? Maybe.

Wrong? Not really.

Gemini and ChatGPT seemed to provide accurate information, but ChatGPT provided the most detailed and helpful information for a user looking to find a plan without the help of a paid trainer.

NOW WHAT?

Now, it’s YOUR turn.

Pick one of the three Generative AI chatbots I’ve mentioned and start playing around with it.

You learn the most by diving in and using the tools.

Below, you’ll also find articles and videos I’ve mentioned so you can do some deeper learning if you’re interested. 👇

REFERENCES

FlowGPT. (n.d.). The Ultimate Library of CHATGPT prompts: Discover, share, and discuss with a vibrant community. https://flowgpt.com/character/famous-people

Gewirtz, D. (2024). How to write better ChatGPT prompts in 5 steps. ZDNet. https://www.zdnet.com/article/how-to-write-better-chatgpt-prompts-in-5-steps/

Google. (n.d.). Google. https://gemini.google.com/?hl=en

OpenAI. (n.d.). OpenAI. https://openai.com/blog/chatgpt

The Royal Institution. (2023, October 12). What is generative AI and how does it work? The Turing Lectures with Mirella Lapata [Video]. YouTube. https://www.youtube.com/watch?v=_6R7Ym6Vy_I