An experiment in AI automation

So, a much as the flood of AI generated nonsense currently ruining the Internet annoys me, that doesn’t mean I don’t think Machine Learning is without uses. And in general, I take the view that you can’t criticise something if you don’t understand it, so I am interested to learn a little of how to use these technologies and what they are capable of.

So, one example I can imagine being useful - I enjoy taking photographs, but I’m terrible at actually titling them or adding keywords. So for this experiment, I’m using an AI to automatically create titles, descriptions, and page keywords when I add a gallery. I leave the AI generated titles/captions here completely unchanged - which means, complete with mistakes - and in a future post, I’ll detail a little of the “how” I did it, and how I’ve made it better (if I have made it better…)

A Knight In A Hat

FUJIFILM X-T5 w/FUJIFILM XF56mmF1.2 R WR (exp 1/450s @ f14/5 (ISO200))
A knight in a helmet is standing in the autumn leaves

A Cat On A Sidewalk

FUJIFILM X-T5 w/FUJIFILM XF56mmF1.2 R WR (exp 1/100s @ f28/5 (ISO200))
A cat sits on a bench on a sidewalk

A Tree In The Forest

FUJIFILM X-T5 w/FUJIFILM XF56mmF1.2 R WR (exp 1/220s @ f28/5 (ISO200))
A red and white chimney in the background

Pumpkins And Flowers

FUJIFILM X-T5 w/FUJIFILM XF56mmF1.2 R WR (exp 1/27s @ f14/5 (ISO200))
A pumpkin sits on a table next to a pot of flowers

3D Printed Flamingo

FUJIFILM X-T5 w/FUJIFILM XF56mmF1.2 R WR (exp 1/10s @ f28/5 (ISO400))
A red flamingo is standing in a forest

A Flower Sculpture In A Park

FUJIFILM X-T5 w/FUJIFILM XF56mmF1.2 R WR (exp 1/58s @ f28/5 (ISO400))
A group of colorful flowers in a park