Images generated by artificial intelligence are once again making headlines. After surging on social networks in recent months, these illustrations of a new genre are beginning to find their place on the platforms of image banks. Shutterstock, rival of the well-known Getty Image, announced a partnership with OpenAI to commercialize images generated by Dalle-E.
“Create at the speed of imagination”
Nicknamed “Create at the speed of imagination“, this program will allow the public “Instantly generate images based on the criteria they enter” on the Shutterstock search engine. To put it more simply, if you do not find images that suit you in the “traditional” bank, Dalle-E will take care of creating one for you.
The partnership between Shutterstock and OpenAI is not very surprising since Dall-E has trained a lot on the photos in the image bank (which have the advantage of often being well captioned). Shutterstock will therefore benefit from the tool it helped to build. But this partnership raises many questions. Many artists are indeed skeptical about the emergence of these image-creating AIs. Formed through the absorption of a considerable amount of digital images, Dall-E, Midjourney, Stable Diffusion and others have made a specialty of imitating the style of certain painters, graphic designers and other artists, often bordering on plagiarism.
Well aware of this problem, Shutterstock has announced that it will set up a fund in parallel with this project which will remunerate the artists whose content has been used to create images on Dall-E. An initiative that highlights the ambiguity that surrounds the development of these AIs. This nest egg will be paid every six months, but no information on the sum has been communicated by Shutterstock. “We recognize that it is our responsibility to participate in this evolution and to ensure that the technology that fuels innovation is based on ethical practices.“, explains Paul Hennessy, CEO of Shutterstock.
The ethical debate rages
Getty Image, for its part, does not see things the same way. The platform has banned AI-generated images for weeks now, and has just given a layer. We “are witness to the recklessness of certain organizations, individuals and companies […] The fact that these questions are not examined is a problem. In some cases, they are simply set aside. I think it’s dangerous and irresponsible. I even think it might be illegal“, blows Craig Peters, CEO of Getty, in an interview with the American media The Verge.
This barely veiled allusion to the practices of its competitor underlines Getty’s difference in approach to these images cobbled together by machines. The American image bank has actually chosen a third way, which consists of using AI to modify images already present on the site. Rather than creating an image from scratch, the artificial intelligence used by Getty is able to alter the skin color, hair or face of certain subjects in the photos in order to “create images that resonate with each audience“.
It’s hard to say if the prospect of turning photos into something else is much more reassuring than that of creating images from scratch. But one thing is certain: we haven’t finished hearing about this subject.
We want to thank the writer of this article for this remarkable content
Shutterstock teams up with Dalle-E to sell AI-generated images
Explore our social media accounts and also other pages related to themhttps://www.ai-magazine.com/related-pages/