Science and technology
An artist-in-residence at Google created a game to test whether people can distinguish between AI and artificial images.
More than a year after the mainstream success of DALL-E 2, the proliferation of AI-generated images is a reality, no longer a hypothesis. Despite the existential handling their emergence prompted, these models are in their early stages and as such continue to suffer from visual problems and inaccuracies.
Perhaps it’s time to learn how to distinguish between algorithmically generated images and human-created images.
A game released by Google on May 3 helps humans hone this very skill. odd one out presents four images to the player. His one of which was generated by Google AI, the player starts with his four lives and gradually gets faster as the game progresses. Categories include cats, cars, candles, and countless others.
in the meantime odd one out The amazing speed of progress of AI image generators has been confirmed. Many of the AI images look indistinguishable from paintings, photographs, and antiques from the Google Arts and Culture collection.
Photo: Screenshot of odd one out.
Start with the background. Anything that looks overly smooth and perfect (pure black or white) or appears in-frame tends to be AI-generated. AI images can sometimes have blemishes or noticeable lighting, but in general, it’s their mediocreness within the group that sets them apart. Finally, Google AI seems to like images that look like patchwork or old prints.
odd one out was created by Caroline Buttet and Emmanuel Durgoni, two artists from the Google Arts & Culture Lab. Their pair also released her three other experiments. XYZToysa word guessing game. undo!, monument guessing game.and imagined haikuvisualize poetry with AI-generated fonts, videos, and music.
Photo: Screenshot of odd one out.
“Our artist residency program has been running since 2014, supporting artists and creative coders who are experimenting with new technologies to solve cultural challenges,” Google Arts & Culture said in a statement. “The starting point for these new experiments was to apply Google AI Image Generation Research to stimulate cultural discovery and learning through play.
Google has released a series of AI-related products in recent months, albeit more cautiously than its technology competitors. This includes some text generation capabilities from within the AI Test Kitchen app and Bard, an AI chatbot similar to his ChatGPT from OpenAI.
Other Trending Stories:
A couple in the UK bought two vases for $10 at a recycling sale. Turns out to be an Art Nouveau collectible worth 150x
Museum renames Van Gogh’s vegetable still life
X-ray scan of Duke Cosimo de’ Medici’s 16th-century Bronzino painting reveals portrait behind mystery
‘He was hungry’: Korean art student stole Maurizio Cattelan’s infamous $150,000 banana from museum wall and ate it
Art Industry News: Rare Blue Diamond Priyanka Chopra Jonas Showcased at Met Gala
A modest collector hid 230 classic cars in a dusty old church.Amazing treasure trove could fetch millions at auction
Christie’s has neglected to reveal the ugly history behind its sensational planned jewelry auction.Then the millionaire’s wife complained
See rare Keith Haring painting on display in Amsterdam for the first time in 30 years – measuring a giant 125 feet
How Lavinia Fontana broke with Renaissance tradition to portray the female nude and became the first female artist to be paid on par with men
Follow Artnet News on Facebook.
Want to stay one step ahead in the art world? Subscribe to our newsletter for breaking news, eye-popping interviews, and sharp critical takes that propel conversations forward.
