#520: People Prefer Ai Art
Midjourney Prompts, Generative AI For Creatives
📣 In this issue:
Bringing game characters to life, people prefer Ai art, meeting the Critterz, Editing with DALL_E, plus Midjourney Office Hours recap.
📰 News-To-Know
1
People prefer AI Art when they think that it has been created by humans. That is according to a study published in Scientific Reports by Simone Grassini and Mika Koivisto reveals an intriguing phenomenon: people tend to have a positive perception of AI-generated art, especially when they believe it's created by humans. The study utilized 20 works from various art genres, including Cubism and Impressionism, created by artists and an equivalent number of works generated by the AI program Midjourney in the same genres. An online survey of 201 participants evaluated their preferences, emotional responses, and perceptions regarding the origin (AI or human) of the artworks.
The findings indicated that participants struggled to accurately identify the origin of the artworks, frequently mistaking AI-generated images for human-created works and vice versa. Notably, there was a general preference for the AI-generated images when participants believed they were created by humans.
2
Wired reports that it is possible to train an AI model without using copyrighted Content
A group of researchers, supported by the French government, unveiled what is considered the largest AI training dataset composed entirely of public domain text, demonstrating that it's feasible to train large language models (LLMs) without infringing on copyright.
3
According to the Hollywood Reporter, the dispute between George Carlin’s estate and podcasters over an unauthorized AI-generated special has been resolved. The conflict arose from a podcast episode created by Will Sasso and Chad Kultgen, hosts of the Dudesy podcast, which used artificial intelligence to mimic the late comedian George Carlin's voice and style. This hour-long special, titled "George Carlin: I’m Glad I’m Dead," utilized AI to generate commentary in Carlin's iconic manner on various contemporary issues, including the rise of reality TV and streaming services.
4
There’s been a significant update to the DALL-E image generation model, allowing users to directly edit AI-generated images within the ChatGPT interface. This new feature aims to streamline the creative process, enabling users to tweak their AI creations without having to leave the chat environment.
You will need to use GPT-4 or invoke the DALL-E GPT first, then work with Chat GPT to define your image. Then just ask it to create the image. You can then select the image and edit it. I got pretty mixed results but it’s early days!