Overview

AI is revolutionising the world of business as we know it, impacting multiple industries by automating tasks, streamlining processes and saving time. ?

In the world of L&D, AI is making a significant impact in the areas of personalised learning paths, automated administration, upskilling and digital content development.?

As an experienced L&D professional, throughout my career I’ve created allsorts of high-quality content ranging from training manuals to elearning modules from interactive videos to cinestories, and whilst looking through my hard drive recently, I stumbled across several Cinestories that I developed some 15 years ago which, got me thinking:

1. Could Gen AI assist me in the creation of a Cinestory?
2. What are the challenges presented by Gen AI when creating a Cinestory?
3. Could I find a solution to the problems, that Gen AI presented?

What is a Cinestory

Cinestories, are great a tool to explain an idea or concept and can be photographic or comic book in style. You can create short stories with 3, 6, 9 or 12 scenes, that impart useful information in a short space of time and are great for learning reinforcement pre - during or post learning events.?

They are ideal for learners who like a quick read that tells a compelling story, or that are located geographically where the internet connection is poor (as the file size is often small). Alternatively, you could use this content to convey content on a particular topic via a Microlearning, as a Cinestory that is written well can be read in just two to three minutes.?

However, in my enthusiastic pursuit to create a Cinestory using Artificial Intelligence, I realised the importance of performing due diligence and research.

A significant hurdle in Generative AI it appears, lays in its' inability to accurately render characters in diverse poses, facial expressions and clothing styles consistently.? ??

Creating a Cinestory - Useful Initial Steps

Before we generate a cinestory, we need the story and the scene script content. For this I'm using Microsoft Copilot (https://copilot.microsoft.com)

Step 1.

Ask yourself what’s the overall learning outcome, the objective. In this example:? ???????? “To streamline my teams use of email and task management using Microsoft Outlook, to enhance department performance.“

Step 2.

Document how many characters are in your story? Let’s say two people.?

Step 3.

What are their names? We’ll say Kate and John for the purposes of this example.??????????????????????????

Step 4.

What are their roles? Kate is an Administrator, and John is a Database Specialist.

Step 5.

Write the main point of your story.

“You are the head of the L&D Department and have identified that a couple of users named Kate & John are struggling with email and task management in Microsoft Outlook, and they need to improve their skills to raise their OTJ task management and performance. You've decided to generate a Cinestory to impart the knowledge they need, and to reinforce content covered in a pending 1:1 coaching / training session“.

Step 6.

Create the prompt that will generate your cinestory:? ??????????????

“You are the head of L&D division and have been tasked with creating a new learning initiative to help staff members named Kate & John, to improve their skills in Task & Email Management using Microsoft Outlook. Create a cinestory which, incorporates a number of learning initiatives that can help Kate & John, improve their skills in the use of Microsoft Outlook."

Step 8.

Review the resulting script and tweak the prompt as necessary.

Step 9.

Decide the GEN AI application you are going to use to create the images that support your scenes. In this example Open Art AI: https://openart.ai/home

Step 10.

Develop the prompt needed to generate the images that support each scene of your Cinestory.

Note: The challenge when using Generative AI is that you might get a great character that you can use one time but what if you see a Gen AI character that looks fantastic, that you want or need to reuse in different positions, with different facial expressions and or clothing. This is where Gen AI has a problem which, various communities I discovered are crying out for a solution to. Nevertheless there are some workarounds I discovered (ChatGPT Community). ?

Prompt Engineering with Open Art AI

Before creating the prompts that will generate the images that support your scenes using the web based Open Art AI application it’s critical that you get really specific about what it is you want.?

Here's a prompt I created that you can experiment with.

Master Prompt

"Create a character in the style of Disney's Pixar, which depicts a business professional working at a computer terminal in a high-rise building. The background will be a simple window view of a cityscape at daytime. Light blue sky with white clouds and buildings in various shapes and sizes filling the background.

The character will be a female business professional with long brown hair, light skin tone, and a business casual outfit.? The outfit the character is wearing will be a white button-up shirt with the top button undone, a grey blazer, and dark grey dress and black dress shoes."

Then add the pose style modification prompt (1, 2, 3, 4, or 5).

Pose Prompts

1. The character is typing: The professional leans slightly forward, typing on the keyboard with both hands.
2. The character is thinking: The professional leans back in their chair, hand on chin, looking thoughtful at the computer screen.
3. The character is talking on phone: The professional holds a phone to their ear with one hand, while the other hand rests on the desk.
4. The character is reaching for coffee: The professional leans forward slightly, reaching for a coffee mug to the side of the computer.
5. The character is stretched and yawning: The professional stands up, stretching their arms above their head and yawning with one hand over their mouth.

Initial Results

I quickly discovered that if you include the Master Prompt shown and combine it with a single Pose Prompt, AI gives you a pretty fantastic result. However, asking AI to develop multiple poses in the same prompt it sometimes gets confused, and explains that it is not possible. This is Gen AI application specific.

I further discovered that, if you apply the above Master Prompt with one pose description prompt and include a prompt that references the Sample: Seed ID No. the results are pretty impressive. (See below - Same Character with different poses and clothing)

Acquiring the Seed ID No.

Step 10.

Acquiring the Seed ID Number in the Open Art AI web app.

Is very strait forward, simply select the image of interest produced by the Open Art AI app, and when the pop window containing both the image and the properties of the file are displayed, highlight the "Sampler: Seed No: 1761030833" for example and incorporate it into your future prompts.

For Example: Use the Master Prompt & Pose Prompts respectively and then add to the bottom a prompt like:

“Generate the character using Sampler: Seed: 1761030833.” A new image is generated identical to the image produced previously.

Step 11.

Once you have acquired the images that support each scene of your Cinestory, launch Microsoft Word and begin creating your Cinestory material.

Upon completion, review and publish the content as a PDF document.

Building Your Cinestory

Build your Cinestory in Microsoft Word by:

1. Using the Add Shapes function with black outline to create the frames.
2. Add Images to Shapes (Right Click Shape Frame | Format Shape | Insert Image ).
3. Search for a Comic Book style Speech Bubble graphic of your choice in *.PNG Format
4. Insert and position the Speech Bubble appropriately.
5. Insert and overlay a text box inside the Speech Bubble.
6. Install Anime ACE Font. ?
7. Return to Speech Bubble and add your script.
8. Review content and then save the document as a PDF ahead of distribution.

I hope you found the content of this article useful, if you require further assistance or have any questions, please get in touch by emailing: [email protected] or contact me directly via WhatsApp: +34 711 071 624.