Q&A with Getty Images CEO Craig Peters on the deal with Nvidia to create AI tools trained on Getty's stock images, the need for AI industry standards, and more
Scott Roxborough / The Hollywood Reporter :
Context & Ripple Effects
Getty’s Nvidia partnership had already produced a licensed-image generative tool and then an iStock service aimed at smaller businesses. This interview places the arrangement within Getty’s wider effort to make licensed visual archives usable in generative AI.
Peters’ emphasis on standards follows Getty’s earlier public focus on copyright, attribution, deepfakes and disinformation in its generative-AI policy discussion. The immediate issue is not only tool development, but the rules governing training data and output use.
First-order effects
- Getty and Nvidia can build AI tools using Getty’s stock-image library, extending a commercial path for licensed visual content in model development.
- Getty’s standards push makes provenance, rights-holder treatment and safeguards central to how it presents AI-generated imagery to customers.
Second-order effects
- Other image libraries and AI developers face greater pressure to offer licensed training access and clearer protections, rather than treating image supply as an undifferentiated input.
- Customers evaluating generated visuals gain a clearer distinction between tools tied to licensed archives and models whose training-data rights are less defined.
Third-order effects
- If licensed-library partnerships become a durable model, valuable content archives may shift from passive licensing catalogs to strategic inputs for AI products.
- Industry standards could become a competitive gatekeeper: providers able to document training rights and output safeguards may be better positioned as legal and customer scrutiny rises.
The trend: Generative-AI vendors are increasingly commercializing licensed content supply and governance assurances alongside model capabilities.