What happens when you ask an AI model to depict different generations? We wanted to find out. With all the talk about generational stereotypes, we wanted to see if AI could give us a fresh perspective on how Baby Boomers, Gen X, Millennials, and Gen Z are perceived. So, we decided to put together an experiment.

We took four models—Stable Diffusion, Midjourney, YandexART, and ERNIE-ViLG—and prepared identical prompts for four generations, focusing on their lifestyles, work habits, relationships, and identities. After running the prompts, we generated 1,200 images and put together our results.

Reflective Boomers, Vibrant Zoomers

Let’s start with the Boomers. We expected the usual laid-back retirement vibes, but what we saw was a little more introspective. Models like Midjourney leaned heavily into this portrayal, often showing Boomers as a bit solemn, bundled up, and gazing off into the distance. These images conveyed emotional heaviness, possibly reflecting a generation weighed down by existential questions.

On the other hand, ERNIE-ViLG took a completely different approach, with 93% of Boomers depicted smiling. This stark contrast between models might be due to differences in the datasets used to train them, highlighting how the model's training and fine-tuning processes can influence output.

Then we looked at Gen Z, the Zoomers. They were by far the most expressive and diverse among all models. We saw detailed, colorful visuals that reflected their individuality and inclusivity. However, an underlying sense of tension and anxiety permeated many of these images, particularly in the Midjourney outputs.

This generation’s portrayals seemed to blend vibrancy with an unmistakable sense of stress, which likely mirrors the pressures Zoomers feel navigating modern life. That said, these images didn’t just show people; they told stories.

Gen X: The Enigma

Gen X? Well, this generation seemed to stump the models. Images of Gen X were less defined compared to other generations. Models also struggled to “age” them correctly, with some individuals appearing too young or too old to be part of this generation. You know the saying that Gen X is the “forgotten” generation?

That’s exactly how they came across—fewer distinct features and less visual flair. The one consistent element? Flannel shirts. Every model zeroed in on that grunge-era vibe, which pretty much solidified Gen X’s rebellious, grunge identity from the 90s.

Gen Y and Gen Z: The Digital Natives

When looking at the images generated for Millennials and Gen Z, we noticed that the models often struggled to distinguish between the two. Both generations were depicted with vibrant, detail-rich images, but the visual differences between them were minimal. This blending of Gen Y and Gen Z into one group made it difficult to identify unique character traits for each generation.

However, there was one consistent difference: Gen Z’s preference for going alcohol-free in social settings. While other generations were often depicted drinking beer, Gen Z stood out for their choice of non-alcoholic beverages, a key divergence from the norm that the models captured.

Stable Diffusion highlighted their work-life balance, with visuals of Millennials hopping from one task to another. The key takeaway? Millennials are seen as adaptable and resourceful, fully embracing modern work culture. But beyond these portrayals, the lack of clear distinctions between the two generations speaks to how AI models may be blending them into a single group.

What Ties Everyone Together? Beer.

Now, for the one thing we didn’t expect. Across all generations, the most common element was…beer. Whether it was Boomers reminiscing, Millennials decompressing after work, or Gen Z in a casual social setting, beer showed up in 34% of the images. It’s the ultimate generational connector—an unintentional yet amusing commonality that speaks volumes about what we might have in common, regardless of age.

What We Learned From the Models

The key takeaway? AI doesn’t just mirror society’s perceptions—it reflects the cultural stereotypes in the data it’s trained on. The differences we saw in how each model represented the generations highlight the importance of understanding what shapes those outputs. While some results aligned with the stereotypes we’re familiar with, others challenged them, offering a fresh lens to view generational identities.

This experiment showed us that AI models have a unique way of interpreting human culture. While generative models can sometimes reinforce stereotypes, they also reveal gaps in representation and biases in the data they’re trained on. These insights remind us of the importance of investigating not only the outputs of AI models but also the datasets that shape them.

Want to see more? Check out the full set of images and dive into the details at AI Generations Project.

Thanks for reading!