PHOTOS: How does AI picture famous Hungarian landmarks?

What if we let artificial intelligence design pictures depicting famous Hungarian landmarks? What would the parliament building or the view from Buda Castle look like according to AI images? Let us show you!

Recently, German artist Boris Eldagsen won the prestigious Sony world photography award, but he refused to accept the prize. His entry was revealed to be a stunt, as he admitted that his image was created by artificial intelligence. His entry was a way to spark debate about AI images and art and let’s be honest, it was a pretty successful idea.

As we have previously asked ChatGPT what an open AI “thinks” of Hungary, in our latest AI experiment, we have tried to see what kind of images an AI text-to-image tool like Gencraft can make about famous Hungarian landmarks and well-known products.

It is worth mentioning that we did not ask the AI to make the images in a particular style, whereas in reality the “artists” can improve their prompts with specific design suggestions, such as creating the images in anime, realistic, watercolour or steampunk style.

Scenery from Gellért Hill

As the renovation of the Citadella and the Gellért Hill hasn’t been finished yet, we’ve asked the AI tool to make an image like it was taken from the top of Gellért Hill. The AI recognises the river Danube and sort of did alright with creating the Hungarian parliament, but likely confused it with Buda Castle. Still, one might be able to recognise Budapest from this picture.

In reality:

Lake Balaton and Tihany Abbey

Next, we asked the tool to create an image of Lake Balaton and the Tihany Abbey. While it did create a very good Balaton scenery, it places the abbey from the top of the peninsula to the lakeside with a very different style, which makes us wonder whether it draws some inspiration from the scenery that can be seen around Lake Como and Garda in Italy.

In reality:

Széchenyi Thermal Bath

In our opinion, the AI did the best job with this prompt, which probably is due to the limited scope of the photos taken inside the bath. The central pool is clearly recognisable, bar the steps, but the colour scheme is also matching what visitors can expect in the summer. It looks like the AI is sticking to geometric shapes, and couldn’t quite grasp the asymmetric structure of the building.

In reality:

Read also: PHOTOS, VIDEO: Budapest’s “skyscraper” opened for visitors, here are the prices

Hungarian Great Plain

Our prompt was to create a picture of the Hungarian Alföld (Great Plain) landscape with a Puli dog, a shepherd and cattle, like how it would look in Hortobágy. Instead of plain, we see a prospering wheat field and a shepherd that does not wear Hungarian clothing. The dog looks like a cross between a Puli and a Bernese Mountain Dog, and the cattle resemble some similarities with Scottish Highland Cattle. Not the most accurate work of the tool.

In reality:

Liberty Bridge with Gellért Hill in the background

Well, this was a hard one. Originally, we wanted to create an image about the Széchenyi Chain Bridge but the tool refused to make images to this prompt. We had to change to the Liberty Bridge, and the tool suffered to make anything that resembles Budapest. The word ‘liberty’ must have confused the tool which could have drawn much more information about New York than Budapest, therefore the bridge resembles the Brooklyn Bridge. The parrot on the fence is questionable as well but could be a fun little addition to the Budapest wildlife.

In reality:

Overall, as these images show photographers shouldn’t worry that AI text-to-image tools will make their job obsolete. However, they are not just fun, in some cases they make surprisingly good renditions of reality based on the information available to them.

Read also: VIDEO, PHOTOS: Astonishing bike path around Lake Balaton ready!