One of my first experiences with generative AI in which I was impressed was back in 2019. Towards the end of 2019, NVIDIA launched StyleGan 2, an updated version of StyleGan. I came across the website thispersondoesnotexist.com. I was surprised to find thousands of realistic photos of human beings, however none of these images were of a real human being. These images were created through generative AI, allowing people to use them freely, without having to find a model. Two years later, I came across MidJourney and ChatGPT. When I was working in a rental platform, similar to AirBnB in which we connect property owners with tenants. I suggested the use of ChatGPT to create creative descriptions of apartments that made them seem more appealing to tenants, increasing the conversion rate. This also saved time and effort for the listing specialist.
When I started using MidJourney, I noticed how powerful it is and couldn’t stop thinking about the future implications this could have on certain industries. Especially when talking about creative industries like the music or production industries. What sparked my curiosity was to what extent does AI generated content belong to you?
The first thing I did was to double check Open AI’s terms and conditions. In the TOS, it is stated that “OpenAI hereby assigns to you all its right, title and interest in and to output”, so according to OpenAI you are the owner of whatever you have used as a prompt and the output. Any image you generate through DALLE belongs to you fully and you have the right to monetize it (OpenAI, 2023). This means that you can use these images in production for commercials, for fillers, to create NFT’s, the possibilities are endless.
Imagine a production company, they could use text to image generative AI tools to optimize their whole process. They can put a model wherever they want, they can get the model to look exactly like they want. Then with a few prompts you could change clothes on the model, or the environment it’s in, or the accessories or objects he/she is holding. The possibilities are endless, significantly reducing production cost for these agencies, hence generating larger profits.
Below are some examples of images I have generated using My Feng Shui that could very likely be used as the advertising for a real brand:
Prompt: A model in the jungle wearing modern tribal clothes in collaboration with a luxury brand
Promt: A model in the beach on a remote island, wearing some crazy sunglasses
Prompt: A male model in some mountains, full of nature, advertising a new phone.
I was able to create these prompts without too much effort. Imagine a production company, hiring prompt engineers, they could greatly benefit from this technology at a very low cost, practically almost for free. The opportunities generative AI has brought to our society are endless.
Citations:
Terms of use. OpenAI. (2023, March 14). https://openai.com/policies/terms-of-use
Hello! I really enjoyed reading your blog article :))! I thought your ideas about ownership and monetization of the pictures made by AI were really intriguing. It was new to me that OpenAI gives you full rights to the product, I wonder why they would want to miss out on such easy profit opportunities. The images you showed us were mind-blowing, although I can still easily tell they were made by AI, i would not be surprised if in a few months these pictures would become so life-like that companies would actually implement your idea. I still do not think it will completely rearrange the modelling and advertising industry, as I think knowing that the person wearing those clothes is an actually existing individual adds a unique human touch making the clothes more appealing. I might be wrong though :)), I guess we will see in a few years/months. PS: The website of the people who do not exist in my opinion is actually a little scary but, wow, very impressive.