
It’s December, that time of year when much of the Western world turns its eyes to Christmas. In that spirit, I’m going to ring out the year with a StrefaTECH series titled “The 12 dAIs of Christmas.”1
For these 12 dAIs of Christmas, each day will have a different theme. There probably won’t be any lords a-leapin’ or golden rings, but I hope to share some resources and ideas to help you in the new year.
On the First dAI of Christmas…
On this first dAI of Christmas, the gift is … Multimedia AI StrefaTECH.
I’ve decided to experiment with AI-generated video using HeyGen, a leading tool for creating video from text scripts. I highlighted it in an early October article, where I demonstrated my mastery of the Italian and Polish languages. (I still find this a bit eerie, but really cool!) Since then, I’ve generated translations in HeyGen into French, so that I could be more welcoming to a high school student who’s from Guinea and is in long-term foster care here in the US; and I created another Italian one that I sent to some friends in Europe, who were duly impressed with the progress of my linguistic skills.2
Today’s Article In Video
Watch this video for today’s article, in which I describe how to use AI to create videos from text that “looks like” it’s you reading the message.
The Script for “AI Deb”
Hi everyone! It’s Deb from strehfuh tek3. I’m delighted that you’re checking out this HeyGen video for today’s article.
First, let’s see a short message from human Deb.
{ video of human Deb }
Thanks, human Deb. We’re back to AI Deb now. And while this video may not be quite right in all ways yet, it’s pretty darned impressive, isn’t it?
Here’s how human Deb created this video.
First. human Deb recorded a 3-minute sample video with her phone. She was seated in front of the fireplace in her living room. The text of that sample video wasn’t important—it was just to allow HeyGen to capture human Deb’s voice, facial expressions, and physical mannerisms, as well as the environment around her.
Second. Deb uploaded the video to HeyGen to create a custom avatar, which is essentially a template of how she sounds and appears.
Third. Deb typed a script of the words you’ve been hearing in this video. She uploaded the script to HeyGen and had it generate a new video—the one you’re watching now. HeyGen uses AI to mimic the style of the uploaded sample video. It uses the background from the sample video (the fireplace, fire, stockings, etc.) and generates an AI Deb who appears to be reading these words.
Finally, human Deb downloaded the AI-generated video from HeyGen and used Canva to combine it with the intro video she had recorded on her phone for this final video you’re watching.
This process isn’t free—it costs about $2 per minute to generate videos. It’s not perfectly polished—yet—though it’s getting better with amazing speed. And it’s probably something that has very limited use in nonprofit organizations.
But for this first day of Christmas, it’s
fun!
Merry, merry Christmas everyone—from both AI Deb and human Deb!
Hey—I heard those groans!
More groans, this time from my husband, who knows how awful my progress has been on languages in the places we’ve stayed over the last two years as vagabonds in Europe!
There are a few words that the AI text-to-audio struggled with, so I followed the recommendation to change the script to spell out a more phonetic version. It did OK, and I’m sure with some further experimentation and tweaking, it could be spot on!