A little over a year ago, Getty Images launched its own Artificial Intelligence (AI) image generator, promoting it as a “commercially safe” option.
Unlike other systems that train on content scraped massively from the web, Getty’s generator is built using their extensive library of licensed content.
This strategy promises greater legal security, protecting users against possible copyright lawsuits. However, it begs the question: Are they really immune to lawsuits or is this just a marketing ploy? Are the generated images of quality and usable, or do they not meet expectations?
We are going to analyze the need for this tool, its operation and how its performance is perceived according to our own experience with it.
The problem of copyright in images generated by AI
Since the explosion of tools such as DALL-E, Stable Diffusion and Midjourney, the use of AI to generate images has sparked a flurry of reactions in the creative industry.
Large companies like Google and OpenAI have been embroiled in legal disputes due to the way they train their models, using data taken from the web without clear licenses.
The lawsuits brought by photographers, artists and media companies underscore growing concerns about the impact of these technologies, not only on copyright, but also on the integrity of creative works.
Getty Images goes on the offensive
Last year, Getty Images, one of the world’s largest providers of photographs and visual content, filed a lawsuit against Stability AI, the company behind Stable Diffusion, alleging that it had used more than 12 million images from its archive without authorization.
This situation not only demonstrates the legal risks that exist, but also highlights a fundamental concern: the lack of control over how content is used in training AI models.
Against this backdrop, Getty and other companies have chosen to develop their own AI tools, trained on licensed material, in an attempt to establish a standard for legal liability in the industry.
Generative AI by Getty: How does it work?
Getty’s AI image generator, called Generative AI by iStock, was developed in collaboration with Nvidia and presented as a safe and reliable alternative for businesses and professionals.
Instead of relying on unverified internet data, this system is trained exclusively on Getty’s vast catalog of licensed images.
This feature not only seeks to eliminate legal concerns but also offer customers an environment where they can create images without fear of copyright infringement.
To get started, the user only needs to register on iStock, Getty’s platform. Once inside, the system allows you to enter a text description of the image you want to create.
The software generates up to four images for each query, and users can define aspects such as the quality, style and format of the final image.
Compensation in case of legal problems
The results are designed to meet the requirements of companies in terms of aesthetics and brand alignment, which aims to make this service attractive to advertising agencies and medium-sized businesses.
In addition, Getty offers a significant advantage in terms of indemnification, in case of legal problems arising from the images generated.
Users who create content with basic prompts are protected up to $10,000 USD, those who opt for an extended subscription can get coverage up to $250,000.
This support is presented as a great attraction compared to other tools that do not offer legal guarantees.
Comparison and opinions: Does Getty keep its promise?
Despite the guarantees and advantages in terms of security, the opinions of users who have tried this tool are mixed. In our experience we compare Generative AI by Getty with tools like DALL-E and Clipdrop.
Although the Getty generator offers adequate and “usable” results, the aesthetic and narrative quality of the generated images was noticeably less interesting compared to other AI generators.
The images tended to be obvious and lacked the atmosphere or visual complexity seen in other AI models.
“Overly literal”
Other users have criticized the tool for its overly literal approach, noting that the images lack the ability to creatively interpret complex prompts.
According to these users, Getty seems to rely on a superficial reading of the prompts, focusing on the basic elements without improvising or enriching the scene, something in which the other generators have made great progress in the last year.
We cannot but agree with this opinion. This behavior could reflect a conservative approach that, while reducing legal risks, limits the creative versatility of the images generated.
The reaction of artists and the visual industry
The arrival of Generative AI by Getty has also had repercussions in the artistic community, which is divided between the need to adapt to these technologies and the fear that their works could be replaced.
However, Getty has implemented some measures to address these concerns. According to statements from its executives, the company will not only pay photographers whose images are used to train the model, but will also offer them a portion of the revenue generated by their subscriptions.
This compensation scheme seeks to balance the participation of content creators, which is a characteristic that Getty shares with Bria AI, a startup that also collaborates with stock image companies.
A new business model or a useless effort?
Getty’s initiative reflects a major shift in the way training AI models is thought about. Faced with an unclear legal framework, companies like Getty, Shutterstock and Picsart are opting for licensing as a strategy to protect themselves and avoid conflicts.
As these companies partner with startups like Bria AI and OpenLicense AI, a business model is taking shape in which content creators can earn compensation for the use of their works in AI, something that could set a new standard in the industry.
Or not. Still, with growing copyright concerns, tools like Getty’s image generator are an important step toward responsible and ethical use of AI in visual content creation.
This post is also available in: