Issue 95: A special how-to issue

Making slides is one of my AI holy grails. I’ve spent many painful hours making slides, and it has never felt like a good use of my time. The slides I need are simple: typically, just explanations of concepts and example sentences.

I’ve tried various slide-making tools before, and none have worked … until now.

Here’s how it went down: When I saw articles about the new Gemini feature today, I was coincidentally reviewing a presentation for my upcoming AP style webinar. I have a completed presentation with more than 200 slides and a fully written script. I don’t need to make a new presentation, but I thought I’d take my script and see what Gemini could do.

The Result: Gemini can quickly make slides that are accurate and look better than the ones I make myself (and I don’t even have a paid Gemini account).

Step 1: Give it instructions

• I uploaded about 1000 words of my script to Gemini as a Word document (I didn’t want to overwhelm it or waste resources by uploading all 12,000+ words).

• I ticked on Canvas.

• I gave it the following prompt:

And I’ll be damned — it made beautiful, accurate slides. Here’s an example of one of the 19 slides it made in about a minute:

Step 2: Make revisions

Revision #1. That first slide might be too small for you to read, and it might be too small for attendees to read, so I asked it to make the text bigger, and it did:

In reality, I might want it even bigger, but I was finished experimenting with that parameter. It can make those kinds of changes.

Revision #2. I asked it to add an arrow to the serial comma in the example, and it did:

Revision #3. It had put a lot of information on some slides, so I asked it to break it up:

And it did. (I didn’t save the original, so I can’t show you a comparison. This is perhaps a flaw in the system — I can’t see a way to go back to earlier versions.)

Revision #4. After breaking up the presentation, I decided I had liked the way one of the earlier slides showed two example sentences side by side, so I asked it to recombine them. I was skeptical that this would work, but it did:

Revision #5. I wondered if I could make smaller line-level edits, so I asked it to remove the quotation marks around the example sentences in the slide above, and it did.

It would likely be faster to just make those small edits by hand though, and you can actually download your presentations to Google Slides and edit all the elements.

A big concern

I am very worried about the images it included in the slide deck. To its credit, it gives you the source for all the images. The downside is that these do not look like images that are free and clear to use. One is from iStockphoto, which costs money if you go to the site. One is from a blog that doesn’t have a photo credit, and you can’t just take photos from random blogs. And the third is from a site called Vecteezy that says it is free with a trial if you go to the page, but I certainly don’t have a trial.

I would never use these images in a presentation. Frankly, I’m shocked Google would include them. I would remove all the images in Google Slides and replace them with images I had the rights to use.

Final Thoughts

Each of the revisions took seconds. I didn’t remake my entire 200+ slide presentation, because I don’t need to, but I originally spent 15 to 25 hours making that deck, and I’m thrilled that I will never spend hours of my life making slides by hand again. (I hope.)

A couple of caveats:

1. I have an extremely complete document from which I’m making slides, so I don’t know how it would work with lower-quality inputs.

2. Sometimes AI systems falter when working on larger projects, so I don’t know how it would work if I tried to make all 200+ slides at once.

Other than that, I was blown away by how fast and easy it was. So blown away that I had to drop everything and tell you about it. 🤓 Based on my past tests, I didn’t have high hopes, but the Gemini-generated presentation was great.

What is AI Sidequest?

Are you interested in the intersection of AI with language, writing, and culture? With maybe a little consumer business thrown in? Then you’re in the right place!

I’m Mignon Fogarty: I’ve been writing about language for almost 20 years and was the chair of media entrepreneurship in the School of Journalism at the University of Nevada, Reno. I became interested in AI back in 2022 when articles about large language models started flooding my Google alerts. AI Sidequest is where I write about stories I find interesting. I hope you find them interesting too.

If you loved the newsletter, share your favorite part on social media and tag me so I can engage! [LinkedIn — Facebook — Mastodon]

Written by a human