With generative artificial intelligence, a phrase can shape extraordinary or hyper-realistic images. Simply describe what you want to create, whether it’s a wooden house with a cotton roof or a stylish office full of busy executives. Then, the tool previously trained on large databases takes the cues and generates never-before-seen images. This powerful resource, which has been in the hands of a few companies, such as OpenAI or Midjourney, is gaining significant presence in the market. Getty Images, a global visual content creator giant, today announced its new generative artificial intelligence (AI) tool to create images or illustrations from a text command. The company argues that it is the first commercially safe in the category: it is trained only with its own image library and with compensation for commercial use.
Getty ensures that the creators of the images incorporated into the algorithm, such as designers or photographers, will receive financial compensation for their contribution to the training data sets. However, the exact amount or percentage of this benefit has not yet been specified. Craig Peters, CEO of the company, details EL PAÍS by videoconference: “It will be a part of every dollar of income that we generate from the service. We will not pay them a flat fee. It is a continuous flow of income for those contributors based on the content they have contributed to us.”
The tool marks a shift in the company’s position regarding generative AI. Earlier this year, Getty filed lawsuits in the United States and the United Kingdom against Stability IA, the company behind the Stable Diffusion image generator, alleging that it used millions of its photos to train its algorithm without permission. In addition, Getty is prohibited from uploading and selling images generated with artificial intelligence artistic tools from other providers on its platform. And according to the CEO, this position will not change and it will not allow images generated through its own tool to be added to existing libraries.
Peters explains that editorial content, such as photographs related to news, sports, entertainment, celebrities or brands, are not included in the new tool’s algorithm, so it will not be possible to generate new images from them. “It’s just our creative content,” he clarifies via video conference.
Getty, whose image bank is used by media outlets, advertising agencies and other companies, is not the first in the sector to incorporate generative AI. In October 2022, Shutterstock added a similar tool to its service and received a lot of criticism from artists whose work has been copied from the web without their consent to train the algorithms.
The new tool will be offered as a standalone service within the Getty platform. This means that both existing and new customers will need to subscribe to it. The price will vary depending on the characteristics of the client, similar to how your business currently operates. By creating and downloading the artificially generated images, users receive a standard royalty-free license that includes representations and warranties, unlimited indemnification and the right to use in perpetuity, worldwide and without exclusivity.
The company guarantees that later this year, customers will be able to customize the tool with their own data to produce images with exclusive style and language. A manufacturer may, for example, generate new visual content based on its own product or brand. However, this feature will not work for photos of people. Currently, the tool is only available in English and, according to what they say, the inclusion of Spanish and other languages will be carried out in successive phases.
From design to photography, the avalanche of artificial content has triggered fear among artists and creatives. Eva Casado, president of the Association of Professional Photographers of Spain (AFPE), explains that the sector is in a “war footing” situation against these generative AI models, mainly because many of them infringe copyright. “You can even order a photograph in the style of a certain photographer. It is a form of plagiarism,” she says. It is also a dispute in the labor market. “We are being affected because there are many jobs that are not being done by photographers right now. AI is doing them because, in addition, you pay a fee per month and create everything you want,” adds the president of AFPE.
For her part, Yolanda Purriños, director of the collegiate body of photography of the Spanish Federation of Image Professionals, emphasizes that those who dedicate themselves to photographing landscapes and nature or advertising “have it very complicated”, given that AI produces images so impressive “that they are impossible to overcome.” “A large majority of creators defend that they are not ‘photographs’, but ‘images’, since they are not created with a camera and cannot be sold as such,” she says.
Due to the constant evolution of photographic equipment, and also the impact of artificial intelligence, Purriños highlights that since the digitalization of photography there has been constant pressure to always stay updated. And now many professionals have returned to the past to regain authenticity. “In this last year the demand for the film has grown [fotográfica]since it is the only thing that maintains the magic and continues to give truth to that captured moment,” he concludes.
You can follow EL PAÍS Technology in Facebook and Twitter or sign up here to receive our weekly newsletter.