Clearly a certain age profile associated with MX5 ownership there 
The AI spits out a picture based on the description fed it.
“Weird old miata owners gsther in the park wearing hats and slurping on a peach”
(You have to be a fan of the band “Presidents of the USA” to get the peaches reference)
“bald men standing around miatas dancing with cats”
“older miata owners wearing shirts and shorts gather in a car park to hop”
As you see, not all generated pictures look possible
“miata and mx5 owners gather in a car park with pet rabbits”
Continuing the ficticious 1960 Miata, I thought “1960 miatas racing at goodwood”
You see its not always matching drivers with cars. But the AI has applied 1960 styling cues to NA MX5, and came up with something plausible, and not unattractive.
Have a play
You need a microsoft email, and probably consent to suck up enormous amounts of data for you. You get credits to generate images quite quickly. When those run out, the images are generated less quickly. Your balence of credits is topped off weekly, and if you are a fool, you can buy credits. Obviously, they’d love you use Bing.
The output is only restricted by your ability to turn imagination into words. There are restrictions; there are certain scenarios that Bing will not generate an image for, eg putting certain living personalities into a scenario. NC Owners; I’m not seeing Bing acknowledge the existance of this variant, except through the occasional windscreen header, as it tries to generate a generic MX5 or Miata.
and not to condone irresponsible behaviour: mx-5 owners cross the forth road bridge drinking cocktails
Ah, the Louisiana-Texas border, where there are drive through cocktail bars (really)
Thanks AST. Highly entertaining! 
“Bagpuss driving a miata made of cake to Janetta’s ice cream shop in St Andrews”
Okay, it hasn’t really captured Bagpuss (this was the best likeness) but otherwise not bad.
I’ve always been fascinated by how MX5s seem to bias heavily toward each end of the age and wealth curves and not so much in between. What other vehicles are like that? BMW saloons, sort of? Anyway, with that in mind.
Average MX5 owners:
Triumph Spitfire meets Lotus Elan meets Frogeye Sprite meets Honda S800.
I’d buy one 
If the questions “average MX5 Owners” are correct, its fascinating the clearly different outputs. Is it also pulling in data from your Internet browsng habits? ie the output says more about you, than the tool.
It’s funny… Someone on Reddit got really defensive over this post. I asked what was wrong with their sense of humour and this was the reply.
"What was supposed to be humorous about your post?
The way actual photographers data was stolen to create these abominations?
The way artists which will never see an ounce of credit or validation for their work getting replicated and reused?
very funny."
They wanted the post to be banned as they were so offended by Ai.
I don’t think they understands how Dall-E works.
"Bing AI Image Generator is a text-to-image diffusion model, which means that it generates images from text descriptions. It is powered by DALL-E 2, a large language model from OpenAI.
To generate an image, Bing AI Image Generator first converts the text prompt into a latent representation. This is a mathematical representation of the text that captures its meaning and structure. The latent representation is then fed into a diffusion model, which generates an image by gradually adding noise to a blank canvas. The diffusion model is trained on a massive dataset of text and image pairs, so it is able to learn the relationship between text and images."