Using AI to Create an Interactive Scenario with the All-New-Adobe Captivate: Part One

This week I will be using AI to create an interactive scenario with the All-New-Adobe Captivate. I’ll do it in several parts because it will definitely be a lengthy one.  I will begin with using AI Based Chat GPT 4 to help generate the content / story structure in this first installation.

Later I will walk you through creating the course structure, the scenario, the interactions, the evaluation mechanics and further leveraging AI to generate the images for the course. I might even explore AI Audio options for music and sound effects.

The star of our little AI adventure will be this kid. And one of our challenges will be getting a consistent set of images that feature our star. But let’s begin as I did, with an idea, and a morning conversation with OpenAI’s Chat GPT 4 engine.

Now we could get terribly technical about what Chat GPT is, but let’s keep this light and I’ll simply say that around 2019 Google made some significant advancements in natural language processing (computers doing a better job of working with language, sentances, writing etc.) with the introduction of Generative Pre-Trained Transformers (GPT.) The advancement in transformers meant that a slightly simplified approach to finding focus for words / concepts among structured language combined with ever expanding parameters to set the stage for massive leaps forward in the results generated by AI to inquiries. Simply put, the computer is able to do a better job understanding words / parts of sentances in the context of … well everything.

So leap forward to today and I’m able to go to OpenAI’s Chat GPT website and have a nice conversation over my morning tea. Here’s what that conversation looked like.

Allen: Good morning.
ChatGPT: Good morning! How can I assist you today?
(You really don’t need to say things like hello or good morning, but it amuses me.)
Allen: I’m working on a little scenario product focused on teaching families safety practices when engaged in recreational activities near a river. I’d like to create a scenario that uses narrative to teach key safety concepts in this regard. The characters would be a father, a mother, a boy, and the boy’s dog.
(Note here that you can be formal or informal, ask it to do things directly or just sort of collaborate, and the tone of that conversation will influence the output. You might say for example that you want to create an outline or a storyboard and get significantly different suggestions. In general I find that being thorough and precise will generate good initial results, and then being specific with feedback will help refine those results quickly.
ChatGPT: Certainly! Crafting a narrative is a compelling way to teach and make a memorable impact. Here’s a short story focused on safety near a river:
Title:The River’s Whisper
The sun painted a golden hue on the river as the Jackson family set out for their annual picnic. The family consisted of father Joe, mother Ellie, their 10-year-old son Tommy, and Tommy’s energetic dog, Buster.
(The writing / introduction of family here is kludgy.  You could rewrite it or try to persuade ChatGPT to be a bit more creative. Think about how you might make the introduction smoother and more engaging – and then consider asking for those changes. At some point however, you would likely have been better off just writing things yourself if you see flaws.)
As they approached the riverbank, Tommy and Buster sprinted towards the water, excited by the shimmering waves. Seeing this, Joe called out, “Tommy! Remember our rules about the river!”
Tommy skidded to a halt and looked back, “Always stay close to you and mom?”
Ellie chimed in, “And never enter the water without letting us know.”
Joe added, “Exactly. Rivers can have strong currents that are not always visible from the surface. We must always respect the river.”
Buster, ever the curious dog, began to chase a butterfly near the water’s edge. As he pounced, he slipped and splashed into the water. Without thinking, Tommy dashed towards the river to save his best friend.
Ellie shouted, “Tommy, stop! Remember, we never rush into the water.”
Tommy froze. He watched as his father, who always wore secure water shoes and knew the river well, carefully entered the water to rescue Buster. Within moments, Joe had Buster safely in his arms, and they made their way back to the shore.
Shaken but unhurt, Buster wagged his tail while Tommy hugged him tightly.
Joe crouched down to Tommy’s level. “Tommy, rivers can be unpredictable. It’s important to stay calm and think before acting. Rushing in might have put both you and Buster in more danger.”
Tommy nodded, looking a little ashamed. “I just wanted to save him.”
Ellie wrapped her arms around Tommy. “We know, sweetie. But safety first. Always.”
The family then set up their picnic a safe distance from the river. As they enjoyed their sandwiches and laughter, the river whispered its timeless song, reminding them of its beauty and power.
Key Safety Concepts:

Always be aware of the river’s potential dangers.
Do not approach or enter the water without informing someone.
Stay calm in emergencies and think before acting.
Equip yourself with the right gear, like water shoes, when near a river.
Always prioritize safety, even in the heat of the moment.

This story can be accompanied by illustrations and can be followed by a discussion or Q&A session to reinforce the safety lessons.

So there are a couple of problems with the first response I got back. First, the response is written in past tense style with an emphasis on writing a story. We really need something that is active and written as the spine of an interactive scenario. In the section below I ask for some modifications to get closer to the desired solution. The most useful aspect of the returned response was the list of key safety concepts. These concepts should probably be at the root of our efforts because they identify appropriate behavior. But we should also try to learn about what people do wrong. Perhaps we should look for a list of things that commonly go wrong related to safety near rivers and moving water.

Back to the conversation

Allen: I like this approach, but since this is an interactive scenario (I forgot to mention that part) could you do something similar, but add key decision points where the main character – let’s make that the boy, has a set of three options to choose from, and then we can see what happens when he chooses well, okay but not perfectly, and when he chooses poorly. Let’s start with just three decisions – that each reinforce key safety concepts.

ChatGPT: Absolutely! I understand better now. An interactive scenario is a great way to engage the users and help them internalize the lessons. Here’s a modified version of the story with three decision points:

The River’s Whisper: Interactive Scenario

Scene 1: The Jackson family arrives at the riverbank. Tommy spots a beautiful, seemingly calm spot on the river with smooth rocks that seem perfect to skip.

Decision Point 1: What should Tommy do?

(Good Choice) Tell his parents and ask if it’s safe to approach the river.
(Okay Choice) Approach slowly and carefully without informing anyone.
(Poor Choice) Rush to the river’s edge excitedly to start skipping rocks.

Outcomes:

For Choice 1: Joe commends Tommy for being cautious and they find a safe spot together.
For Choice 2: Tommy almost slips on a wet stone but manages to steady himself.
For Choice 3: Tommy slips on a mossy rock and falls, getting wet and scraped up. Ellie reminds him of the unpredictability of riverbanks.

Scene 2: After their picnic, Tommy sees some children playing with a ball that accidentally falls into the river, quickly drifting away.

Decision Point 2: How should Tommy react?

(Good Choice) Alert the other children’s parents and stay put.
(Okay Choice) Look for a long stick or something to retrieve the ball without entering the water.
(Poor Choice) Jump into the water to grab the ball.

Outcomes:

For Choice 1: The parents thank Tommy and decide it’s better to let the ball go than risk safety.
For Choice 2: Tommy tries with the stick but can’t reach the ball. At least he tried safely.
For Choice 3: Tommy struggles against an unexpected current. Joe rushes to help him out, reminding him of the hidden dangers of rivers.

Scene 3: As the family packs up, Buster starts barking at some ducks, teasing him from the middle of the river. Eagerly, he looks like he wants to swim over.

Decision Point 3: What should Tommy do?

(Good Choice) Call Buster back using his favorite treat or toy as a distraction.
(Okay Choice) Shout for his parents but make no move to approach the water.
(Poor Choice) Try to enter the water to grab Buster before he can chase the ducks.

Outcomes:

For Choice 1: Buster comes running back, wagging his tail, and the ducks swim away.
For Choice 2: Ellie comes over with a treat and helps Tommy call Buster back.
For Choice 3: Both Tommy and Buster end up in the water, causing panic. Joe must help them both out, reiterating the importance of safety.

Now we have a better, more useful outcome. The language / description is now active, but we don’t know that we’ve selected the right objectives for the course. Do these relate to the kinds of things that cause actual harm to people (especially children) near fast moving water bodies?

Some of this information can be retrieved from Chat GPT, but best practice is to do the research as well to ensure both that your research is up to date, and that the results are not flawed.  Here’s what it looked like when I probed ChatGPT to find out how the initial decision points related to the actual dangers faced by children near large bodies of rapidly moving water.

It’s worth noting that this – our problems, is where we should begin assembling most any scenario. This is basically a roadmap to the behavior that we want to change. Things like accidental slips, underestimating the strength of the current and distractions – like pets are the elements most commonly associated with accidents related to rivers. So these are the behaviors that we should seek to improve. Each one will have it’s own best practices, and in a scenario we want to design interactions that give learners a chance to discover (without facing any immediate peril) what it might be like to make different choices when faced with these challenges.

I also wanted to understand what, if anything, ChatGPT knows about things like this. What are it’s information sources and how reliable are they? Is this just the result of reading fan fiction, or are there some actual sources of data for these assertions. (Keeping in mind that computers don’t ‘understand’ any of this, it is rather a question of probability analysis to predict relationships – meaning that one should ALWAYS verify facts and information – because those probability driven relationships can and do result in unexpected replies sometimes.)

Here you can see that as of 2021 ChatGPT did have well resourced information regarding drowing and other accidents and was using reputable sources to discover the data. The results here were quite good and it even advised me to consult the best resources for updated information from the WHO and other valid sources.

Now we are getting somewhere. We have a clearly defined and understood problem.

Too many children are injured or killed near large bodies of rapidly moving water. The most frequent cause of those accidents are riverbank falls, underestimating strength of current and lack of supervision. Our scenario should present best, mediocre and worst path choices for a learner who will make choices on behalf of a child in the story. We will provide unlimited attempts to achieve the best outcome, and will present the choices in a way that allows learners to safely visualize the potential outcomes of each choice. Our goal will be to help learners make better decisions around rapidly moving bodies of water, and ultimately to save lives as a result.

In the next blog I will create a decision map and start creating the specific dialog and images for the scenario in the All-New Adobe Captivate itself. As usual, let me know your thoughts and feedback in the comment section below.

I feel like Buster may be heading for trouble.

All the little things you might want to know:
1. Here’s a link to an article about the two new versions of Adobe Captivate and how they differ.
https://elearning.adobe.com/2023/07/two-new-adobe-captivates-what-does-that-mean/
2. Here’s a link to download the All-New Adobe Captivate (it’s a trial but it will activate if you login and you already have a subscription to Captivate.)
https://www.adobe.com/products/captivate/download-trial/try.html?sdid=CMBLG
(Sign in with ‘Personal Account’ instance of your Adobe ID if you receive “Account Type not supported” error)

The post Using AI to Create an Interactive Scenario with the All-New-Adobe Captivate: Part One appeared first on eLearning.

Similar Posts

Leave a Reply

Your email address will not be published. Required fields are marked *