I asked AI to make images of doulas. Here’s what it gave me.

Maybe it’s just because I’m in academia, and am married to someone who is very intrigued by Chat GPT, but it feels like AI-created content comes up in half of my conversations these days. Despite that, I hadn’t really experimented much with it myself. (I have a dissertation to finish, after all!)

But then, a week ago, as I lay awake in the middle of the night, I decided to see what AI thinks doula support looks like.

It was, well, strange.

This foray began as a curious exploration (with, I will admit, a lot of suspicion and interest in the scoffing potential of what I would find). It soon became a kind of gnawing, almost haunting experience with some beautiful moments and some odd trends.

First, AI thinks a lot of what I do as a doula involves touching people’s bellies.

Another theme was touching people’s chests.

I am happy to report that despite the fact that I proudly keep my certification in First Aid and CPR, I’ve never found myself in these types of positions.

It got weirder. And I got kind of worried about AI’s mental state when I looked up “doula support” and got these images:

If a doula is making you feel as despondent as those people look, GET A NEW DOULA.

But that’s not all that I thought was off about these images.


AI is falling into the same trap that countless medical textbooks and pregnancy guides have created and furthered.

Check out the results when you just describe your ideal art as “doula”:

Notice how no one has a face?

AI probably struggles to create realistic-looking faces, so maybe portraying everyone without faces is a way to avoid creating images that look less realistic. But it can dehumanize someone to depict them as their body parts (such as a stomach) rather than their full selves. (Here’s an article that talks about how some pregnant women who participated in a photovoice study saw themselves as just bellies, and here’s an article that talks about a broader problem of representation in medical textbooks.)

So I’ve been actively trying to move away from photos of pregnant bellies or diagrams that show cross-sections of the human body while ignoring the wholeness of how our parts work together (especially in childbirth education class). Because one of the below photos really highlights a sense of connection and human dignity to me, and the other three do not.

How do you prioritize respect for the whole human person in your profession? Or do you have ideas on how I can work to do so in mine? Comment below!

Third, AI thinks doulas are about babies.

Clip art babies.

CGI-looking babies.

And pixel art babies.

I love babies.

But, like most doulas I know, I became a doula, not to hold babies, but to hold parents.

That being said, I really liked the following images of babies:

Fourth, giving the algorithm different styles got me closer to getting images of real doula work.

We started with images like this:

And this:

To me, these don’t at all portray the experience of having a doula. My clients describe me, and I describe the doula I had at my daughter’s birth, as warm, caring, supportive, engaging…not exactly the words that came to mind when I saw the first images AI created for me.

But then I started seeing some images that seemed more like what birth support usually looks like. Proximity, presence, comfort. Bearing witness. Holding.

Then we went back to touching bellies.

But these images were giving more of a sense of the gentle comfort and holding that I strive to provide as a doula.

Also, I noticed that birth started to get more dynamic the more images I requested. I even started to see birthing people not lying flat on their backs!

Then, I just started having a lot of fun with different styles.

I think some of these are really beautiful. After reflecting on this, I realized that the reason I’m drawn to several of them (even if they focus on babies, or show doulas providing counter pressure like it’s CPR) is because they highlight the main concept of my small business: holding. Like I say in my bio, that’s what I do. I hold. From the most exciting of outcomes to the most heartbreaking of outcomes, I hold – space, conversations, energy, emotions, faith, doubt, and most of all parents. I treasure the whole person with my whole heart, no matter age, race, gender, creed, or orientation. Through my training, experience, and creativity, I provide or find the calm and the resources families need in tender moments of transition, and I companion them as they move through those moments.

Some of these photos really show that to me, and I hope you see and feel it too.

And some of these are just kind of weird. But I hope you find those entertaining.

Please enjoy.

Styles based on Lego and Mary Cassatt both have a lot of lying on your back, which as a childbirth educator I have to say is not my favorite thing. On the other hand, these images have (for me, at least) a beautiful nostalgia.

Check this Picasso style out! That’s not my typical favorite art style, but I’m really intrigued by these depictions.

On this style, “children’s book illustration,” I liked the first picture and then just got more confused with each subsequent picture.

I don’t really know what I thought I would get with “stained glass” as the style, but it wasn’t this.

And now, I give you the winner of the best depiction of hip squeezes award: Vermeer style.

And, lastly, because you know I had to:

(I’m a little surprised at all the red shirts, to be honest. Seems like maybe not the outfit AI would want to put people in.)

Now, I want to share with you some of the styles I found most touching.

All in all,

AI seems to have some misconceptions of what doulas are and do, but some of what it produces does reflect aspects of what doula support can be like.

And a last note: I won’t be using AI images for branding anytime soon.

Do I look much like that fellow on the right, there?

How about this gal?

I think I’ll stick with my current headshot.

Weigh in!

Have you used Dall E or other AI software to create images of your profession?
What other styles should I try?
What’s your favorite and least favorite image here?
Tell me everything by commenting below!

1 thought on “I asked AI to make images of doulas. Here’s what it gave me.”

  1. I’ve never tried image AIs, but have been having a field day with ChatGPT. As a writer, that stuff is incredible. One thing I’ve noticed about the images you’ve prompted though, is that AI has the same problem with humans—drawing hands. The hands in those images are just freaky, lol.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s