Getty and Nvidia partner to integrate generative AI with stock photography
January 8 2024
Getty Images and Nvidia are expanding their AI collaboration by launching “Generative AI by iStock,” a new text-to-image platform designed to create stock photos on demand. Unlike its predecessor, Generative AI by Getty Images, which caters to a multiuser enterprise audience, iStock’s version is tailored for single users or small and medium businesses, offering more workflow efficiency and tailored imagery without needing physical photoshoots. The AI, powered by Nvidia’s Picasso model and trained exclusively on Getty and iStock photo libraries, omits training on editorial images to avoid generating known trademarks or personalities. At a cost of $14.99 for 100 prompts, the service presents features such as Inpainting and Outpainting for image refinement, but it offers limited legal indemnity up to $10,000 per asset, unlike the unlimited coverage from Getty’s initial AI service. Users can also engage in revenue sharing through Getty’s existing licensing model.
Back to Breaking AI News
What does it mean?
- Generative AI: A type of artificial intelligence that is capable of generating new content, such as images or text, based on its training data.
- Text-to-image platform: A system that can create images from textual descriptions using artificial intelligence algorithms.
- Multiuser enterprise audience: Typically refers to software or services designed to be used by organizations with multiple users, such as companies or large businesses.
- Workflow efficiency: The optimization and effectiveness of a series of steps or tasks that comprise the work process within a business or project.
- Tailored imagery: Custom-made or specifically designed images that meet the particular needs or requirements of users.
- Physical photoshoots: The act of taking photographs in a real-world setting, often involving models, photographers, and equipment.
- Nvidia: A technology company known for its graphics processing units (GPUs) for the gaming and professional markets, as well as its system on a chip units (SoCs) for the mobile computing and automotive market.
- Picasso model: A specific artificial intelligence model developed by Nvidia, named here as 'Picasso', though this may refer to an internal or project-specific name as the name is not widely recognized outside the given text.
- Editorial images: Photographs that are often used in newspapers, magazines, or online publications to illustrate stories and articles.
- Trademarks: Distinctive signs or symbols used by a company to identify and distinguish its products or services from those of others.
- Inpainting: A technique in image editing where the AI fills in missing or damaged areas of a picture.
- Outpainting: An AI technique used to extend the borders of an image and create a coherent continuation of the scene beyond its original edges.
- Legal indemnity: A legal guarantee or protection against possible loss or financial burden, here referring to a limit on how much the service provider would cover in case of legal issues with the asset.
- Revenue sharing: A financial arrangement where revenue is shared between parties, in this case, between the user and Getty based on licensing models for the images created.
- Licensing model: A business model that allows for the authorized use of a company's products or services, which in this context refers to the terms under which Getty Images allows its content to be used.