Technology and Creativity: Advances and Perspectives in Artificial Intelligence
Posted: Tue Dec 24, 2024 7:13 am
During the Nvidia 2024 Global Conference, a fascinating talk addressed the intersection of artificial intelligence and human creativity. Using Getty Images as an example, the company discussed how it developed its own Generative AI, based on Nvidia’s Picasso solution, to generate copyright-safe images that earn fair remuneration for creators.
In the talk, Andrea Gagliano, Senior Director, AI/ML at Getty Images, discussed how the platform continually compensates photographers for the photos used in training the tool, in order to encourage the creative industry to continue producing materials. Andrea comments that, even with an image bank that represents only a fraction of what other market solutions like Midjourney use, they used all of Getty's experience to create a platform that delivers superior results for commercial uses.
Getty’s Generative AI was also designed much more intentionally than others to more fairly represent society and provide more natural, higher-quality results. So when you enter a prompt about a doctor into Getty’s AI, cameroon whatsapp number database get results that are less biased than you would get from other tools. And when you input something like a parent changing a diaper, you get more natural, non-fanciful results.
During the Nvidia 2024 Global Conference, a fascinating talk addressed the intersection of artificial intelligence and human creativity. Using Getty Images as an example, the company discussed how the company developed its own Generative AI, based on Nvidia’s Picasso solution, to generate copyright-safe images that are fairly remunerated for creators. In the talk, Andrea Gagliano, Senior Director, AI/ML at Getty Images, discussed how the platform continually compensates photographers for the photos used to train the tool, in order to encourage the creative industry to continue producing material. Andrea said that even with an image bank that represents only a fraction of what other market solutions like Midjourney use, they have leveraged Getty’s expertise to create a platform that delivers superior results for commercial use. Getty’s Generative AI was also developed with much more intentionality than others, in order to more fairly represent society and deliver more natural, higher-quality results. So when you enter a prompt about a doctor into Getty’s AI, you get results that are less biased than other tools. And when you input something like parents changing diapers, you get more natural, non-fanciful results.
Andrea also emphasizes that the decision to pay the people responsible for the data used in training should also be seen as a question of sustainability of databases for creative training. After all, if there is not enough data for the model to evolve, it will become increasingly incorrect.
As an example, she brought up the prompt of four generations coming together, which could become increasingly common in the future, with people living close to 100 years. Today, there are still not enough images of this scene for the model to understand the demand. And this prevents it from responding efficiently. In other words, if creatives stop reflecting the world and recording social changes in large volumes, the models will stagnate.
AI and creativity at Nvidia's GTC event
GenIA can't create images of four-generation families because there isn't enough data
To reinforce this point, Andrea draws an interesting parallel between the evolution of the most searched images on Getty Images in recent decades, showing how visual records reflect social changes. According to Andrea, when we think about remunerating creators, we are not just thinking about them or the market as it was, but rather about the future and the sustainability of Generative AI models themselves.
Transformers and AI: panel with the pioneers of current AI models
AI and Creativity at Nvidia's GTC
The informal panel brought together, for the first time on the same stage, the inventors of the Transformer model, the architecture that made possible the emergence of tools such as GPT Chat.
We were able to follow the story behind the creation told by the pioneers of the invention, from when the idea came about to solve problems with search speed and translation on Google to curiosities such as the model almost being baptized with the unfriendly name of CargoNET.
Jensen Huang, CEO of Nvidia, reinforced the importance of the model for the evolution of Artificial Intelligence and stimulated the debate about the next frontiers for the evolution of the models. They commented how the next step will probably be to have something capable of recognizing the size of the challenge and using the necessary computing power in order to make AI more efficient.
In the talk, Andrea Gagliano, Senior Director, AI/ML at Getty Images, discussed how the platform continually compensates photographers for the photos used in training the tool, in order to encourage the creative industry to continue producing materials. Andrea comments that, even with an image bank that represents only a fraction of what other market solutions like Midjourney use, they used all of Getty's experience to create a platform that delivers superior results for commercial uses.
Getty’s Generative AI was also designed much more intentionally than others to more fairly represent society and provide more natural, higher-quality results. So when you enter a prompt about a doctor into Getty’s AI, cameroon whatsapp number database get results that are less biased than you would get from other tools. And when you input something like a parent changing a diaper, you get more natural, non-fanciful results.
During the Nvidia 2024 Global Conference, a fascinating talk addressed the intersection of artificial intelligence and human creativity. Using Getty Images as an example, the company discussed how the company developed its own Generative AI, based on Nvidia’s Picasso solution, to generate copyright-safe images that are fairly remunerated for creators. In the talk, Andrea Gagliano, Senior Director, AI/ML at Getty Images, discussed how the platform continually compensates photographers for the photos used to train the tool, in order to encourage the creative industry to continue producing material. Andrea said that even with an image bank that represents only a fraction of what other market solutions like Midjourney use, they have leveraged Getty’s expertise to create a platform that delivers superior results for commercial use. Getty’s Generative AI was also developed with much more intentionality than others, in order to more fairly represent society and deliver more natural, higher-quality results. So when you enter a prompt about a doctor into Getty’s AI, you get results that are less biased than other tools. And when you input something like parents changing diapers, you get more natural, non-fanciful results.
Andrea also emphasizes that the decision to pay the people responsible for the data used in training should also be seen as a question of sustainability of databases for creative training. After all, if there is not enough data for the model to evolve, it will become increasingly incorrect.
As an example, she brought up the prompt of four generations coming together, which could become increasingly common in the future, with people living close to 100 years. Today, there are still not enough images of this scene for the model to understand the demand. And this prevents it from responding efficiently. In other words, if creatives stop reflecting the world and recording social changes in large volumes, the models will stagnate.
AI and creativity at Nvidia's GTC event
GenIA can't create images of four-generation families because there isn't enough data
To reinforce this point, Andrea draws an interesting parallel between the evolution of the most searched images on Getty Images in recent decades, showing how visual records reflect social changes. According to Andrea, when we think about remunerating creators, we are not just thinking about them or the market as it was, but rather about the future and the sustainability of Generative AI models themselves.
Transformers and AI: panel with the pioneers of current AI models
AI and Creativity at Nvidia's GTC
The informal panel brought together, for the first time on the same stage, the inventors of the Transformer model, the architecture that made possible the emergence of tools such as GPT Chat.
We were able to follow the story behind the creation told by the pioneers of the invention, from when the idea came about to solve problems with search speed and translation on Google to curiosities such as the model almost being baptized with the unfriendly name of CargoNET.
Jensen Huang, CEO of Nvidia, reinforced the importance of the model for the evolution of Artificial Intelligence and stimulated the debate about the next frontiers for the evolution of the models. They commented how the next step will probably be to have something capable of recognizing the size of the challenge and using the necessary computing power in order to make AI more efficient.