Getty Images introduced Generative AI through iStock earlier this year, providing small businesses, content creators, and advertisers with access to cost-effective and secure conceptual AI capabilities. This innovative tool aims to enhance creativity by enabling clients to quickly and affordably create customized photos while minimizing risks. The launch of this new technology has sparked curiosity, particularly regarding its potential impact. In a recent discussion with Grant Farhall, Chief Product Officer, and Bill Bon, Senior Director of Creative Operations at Getty Images, I delved into their insights and explored the expanding capabilities of this platform.
The element of chance plays a significant role in the realm of AI, especially in the context of general AI. Many large language models are trained on data that may lack sufficient intellectual property rights, raising concerns about legality. Therefore, the assertion by Grant Farhall that Getty Images’ AI systems are both economically safe and constitutionally protected is intriguing. The question arises: how is this level of security achieved?
Getty Images emphasizes that its conceptual AI model is derived solely from its vast creative library in collaboration with NVIDIA. By maintaining full control over the training dataset and understanding the model’s training process, Getty Images ensures the integrity of its AI systems. Grant Farhall underscores the importance of avoiding any potential infringement on intellectual property by carefully curating the training data. This meticulous approach provides users with the confidence to experiment with AI-generated content without fear of legal repercussions, as Getty Images guarantees constitutional protection for iStock-generated images, with additional coverage options available for other platforms.
This strategy hinges on the premise that quality inputs yield exceptional outputs. Leveraging its extensive collection of creative assets, Getty Images positions itself at the forefront of AI-generated content creation. The diversity within its library enables the AI model to generate a wide array of images, showcasing its versatility. However, certain sensitive topics such as prominent personalities, landmarks, and brand logos are deliberately excluded from the training data to mitigate risks. This precautionary measure, coupled with technical safeguards, ensures that potentially problematic content prompts are promptly identified and addressed.
Getty Images remains committed to upholding the integrity of its AI model by refraining from claiming ownership of the generated content. Despite utilizing the suggestions and images to refine the model, Getty Images does not incorporate the generated graphics into its creative library for general use. This approach underscores Getty Images’ dedication to maintaining transparency and safeguarding against unintended consequences arising from AI-generated outputs.
Ultimately, Getty Images aims to empower its clients to leverage AI technology responsibly, providing reassurance that legal recourse is available if needed. Through a combination of stringent quality control measures and proactive risk management strategies, Getty Images strives to instill confidence in users while fostering a culture of innovation and creativity.
