Artificially Generated Artwork wins First Prize in Competition


The fuss around so-called Non-Fungible Tokens or NFT’s seems to have died down this year, as the wider crypto market continues so suffer huge losses with no clear sign of when things might turn around. One of the most criticized aspects of the NFT craze was the fact that many of these “artworks” were generated by artificial intelligence algorithms, requiring little to no effort on the part of an actual artist.

Now, a man has used such software to produce a virtual painting that he entered into a competition and was shocked to find that he was awarded the first-place prize in the contest. The image, titled “Théâtre D’opéra Spatial”, appears to be an impressive painting of an opera scene with performers on stage, yet it was created by a text-to-image generator known as Midjourey. The man was awarded the first prize of $300 after submitting a printed version of his image to the United States Colorado State Fair’s fine art competition, before later admitting the source of the image – much to the chagrin of many observers.

Artificially Generated Artworks

It’s an undeniably impressive feat on the part of the man, who said he used thousands of written descriptions which he fed into the text-to-image generator in order to create his prize-winning piece. Many people have expressed their interest in using the technology to help with somewhat mundane computer graphics tasks, such as generating non-playing characters for use in video games or creating low-paying symbols for the best slots sites. Such tasks usually require lots of repetitive work from a skilled individual, whose time could undoubtedly be better spent on other things. Software such as Midjourey could enable such people to use their time more efficiently, supporters argue.

The man who created the image stated that he knew it would be controversial after he revealed the source of his artwork. He noted that many people who had previously derided AI-generated artworks on twitter had praised his picture before it was revealed to have been computer-generated and expressed his distaste towards the fact that these same people had since deleted their messages of support, which he believes proves that they are not analyzing artworks based on their visual value.

What Next for Artificial Intelligence?

Google’s “DeepMind” division has already trained robotic players to play football, according to some reports, using reinforcement learning to control motor and teamwork skills amongst a group of several robotic players. The company claims that Football is an ideal game for testing its software’s planning skills in the physical domain, as it requires the bots to learn how to move and co-ordinate their parts at the same time as working with others towards achieving a goal.

A demonstration of this technology in action would be a huge leap forward for artificial intelligence technology, which only managed to master the board game chess three decades ago. To see a group of machines participating in a more physically demanding sport such as soccer would undoubtedly mark a new era in the development of artificial intelligence and neural networks.

On a lighter note, a new app claims to be able to decode the meaning of your cat’s “meows” into English, using machine-learning algorithms. The app is fittingly named MeowTalk and is already available in the app stores of both Android and Apple platforms, although users have reported mixed success in its readings of how their cat is feeling. The question is, do the owners genuinely know their animals better than the algorithm does?



Source link

Comments are closed.