My latest Generative AI inspired class presentation creation process

I teach instructional design courses in an online masters program. I’ve been teaching for more than eight years now so some of my slide presentation content – which used to be in VoiceThread – needs to be updated. This semester I’ve embraces some Generative AI tools to help enhance my presentations, which I now provide as short videos (about 5 minutes in length) to my students. Here’s a breakdown of my latest process.

1. Start with an Outline

I always begin with an outline of what I want to communicate. Since I’ve done similar presentations in previous years, this gives me a foundation to build on. The outline helps identify the key concepts and the flow of the presentation. It’s the skeleton of my slides—quick bullet points that will later flesh out into full ideas. Fortunately I get a lot of this content from my previous slides.

2. Break Down Concepts for Clarity

Once I have my outline, I revisit how I presented the concepts before. If I was too concise in past presentations, I break down the ideas further, spreading them across more slides. This step ensures that each key concept has room to breathe and the audience can grasp the material at a comfortable pace.

3. Script Enhancement with ChatGPT

After determining the concepts, I take my initial outline or text and ask ChatGPT to transform it into a refined script. My focus is on the key ideas and ensuring the language is clear, engaging, and tailored for the audience. ChatGPT helps to enhance the flow and style, making sure that the narrative is compelling. I’ve done it with this blog post as well!

4. Editing and Instructor Notes

Once I get the enhanced script back from ChatGPT, I review and edit it to match my voice and ensure the message is on point. I place the edited scripts in the instructor notes for each slide.

5. Designing Key Points & Images for Slides

On each slide, I add a visual and sometimes one to three key points of text. For creating the visuals, I use a four different tools depending on what the visual is about.

• ChatGPT for Image Ideas and DALLE for first take on images I first ask ChatGPT for ideas on what type of images would best match the slide’s script. After settling on an idea, I ask ChatGPT to help craft a prompt that I can use to generate the image. I ask it to generate the image and see if that aligns with what I am looking for.
• Adobe Firefly for Visual Creativity If ChatGPT doesn’t give me what I want, I try Adobe Firefly to take my image further. While similar to ChatGPT, Firefly gives me more creative control. By uploading reference images or inspirations, I can keep the visual style consistent across slides.
• Creating Diagrams with PowerPoint If the slide needs a diagram or chart, I turn to PowerPoint. Its SmartArt feature is excellent for crafting clean and professional diagrams such as relationship diagrams and process flows.
• Vector Art with Adobe Illustrator For more tailored visuals or icons, I use Adobe Illustrator. It’s particularly useful if I want vector images that are easily editable. I can generate a simple image and then make minor tweaks or different iterations. Illustrator does a better job of making successive iterations of a given generated image than the other tools I’ve tried.

6. Export Slides to PNG

Once the visuals and text are finalized, I export each slide as a PNG file.

7. Generating Audio with Elevenlabs.io

I use Elevenlabs.io, which offers a voice clone feature, to do the audio narration. I input the script for each slide separately—doing one slide at a time to maintain audio quality. Longer texts can affect performance, so I keep it bite-sized for clarity and coherence. I listen to each slide to make sure it is close to natural. I still need to ask my students for feedback on the slides to see what they think of the audio narration.

8. Slide & Audio Assembly in Camtasia or Premiere Pro

With both the slide PNGs and audio files ready, I use either Camtasia or Premiere Pro to put them together. Each slide gets its corresponding audio track, and this step is all about ensuring smooth transitions and a cohesive flow.

9. Audio Enhancement again using AI

I then use the auto-fix audio tool to enhance the sound quality. This makes sure the audio is crisp, clear, and professional. Adobe Premiere pro does a better job of this than Camtasia.

10. Export Video as MP3

Once the slides and audio are perfectly synced, I export the video as an MP3 file.

11. Import into the Learning Management System (LMS)

Finally, I import the finished product into the LMS. At uMass Boston we use Echo360 which does a great job of creating close captions and transcipts. One of the advantages of using the AI clones audio is that the auto captions are higher quality such that I rarely need to make edits or manually upload transcripts. At this point, the presentation is polished, and ready for students or audience members to enjoy.

I have used Premiere Pro to generate captions and then export the caption file and import it into Echo360, but I find that is an extra step that isn’t necessary as the auto captions are high quality.

That’s my process.

What do you think? Is this a good use of generative AI for instructors?

** ChatGPT used to edit a draft of this post and Adobe Firefly used to create the feature image.