It’s 2023, and Artificial Intelligence is truly upon us.
In recent months, the world has been taken by storm by OpenAI’s much-anticipated launch of their new chat interface for their AI systems, called ChatGPT. The system opens up access to AI to everyone, for free, in an easy-to-use interface that users are instantly able to adapt to, given its similarity to chatbots and instant messaging. Access to OpenAI’s products has been available previously to those “in the know” but that access required technical expertise not usually possessed by the average Joe.
Anyone with an internet-connected device can now ask an AI questions, and be able to ask follow-up questions too, to increase the quality of the feedback, and this is handy for a huge number of use cases. This unprecedented access to “curated” knowledge, beats out the tried and true method of using search engines to hunt down and compile facts. What used to take hours of research can now be accomplished in seconds with a carefully worded query to the AI. It’s being used to write blogs, compose emails, write poems, proofread content, write essays, formulate business plans, and the list goes on and on. The large language model (LLM) data that ChatGPT has been trained on ends in October of 2021, and it’s not currently connected to the internet, but there are other types of AI out now in the wild that have been rather more controversial. But this isn’t the only way that AI is being used.
A new era of AI-generated art has also dawned on the world over the last two years, with the launch of DALL-E, an AI image generator based on the principle of Stable Diffusion. DALL-E is also owned by OpenAI, and uses a gigantic data set of images scraped from all over the internet to make an educated guess about what an image should look like when given a simple text prompt by a user. Since DALL-E was released, other services have sprung up offering similar results, including Midjourney, which is currently thought to be one of the best AI art generators. Within seconds of sending a prompt, users can receive multiple detailed images in any art style under the sun, and this is where AI art has begun to become a bit unstuck.
The data set, LAION-5B, which these systems use, contains over 5 billion images, and these images include huge swathes of artists’ copyrighted work. The AI is able to distill the essence of the artists’ style and regurgitate it in new formats on demand. You could ask for a picture of a dog, wearing a hat, in the style of the Mona Lisa, and you’d get back something very much resembling that request. Artists around the globe have been speaking out about the use of their artwork in the dataset without their permission, and there are numerous lawsuits currently popping up in which AI companies, including Midjourney and OpenAI, are being sued for copyright infringement.
An AI generated image of a dog wearing a hat, in the style of the Mona Lisa.
Copyright lawsuits between musicians often argue that their melody, rhythm, or lyrics have been unfairly copied, and their songs are dissected to determine which elements are too similar and appear to have been “stolen”. AI art is less clear. The output is clearly viewable, but there’s nothing in the results to say which images the AI pulled in to formulate this new artwork. Without being able to trace the source of the “inspiration” it’s very difficult to prove a copyright violation. Some AI-generated images have been found to have faint hints of an artist’s signature in them, which the AI thought was a feature of the art, rather than a watermark.
The result of those court cases will be pivotal in how we as humans continue to use and develop AI in our everyday lives. Once a precedent is set in one courthouse, it will have global repercussions which could affect the very future of art, and the livelihoods and rights of artists. The case will be won or lost on whether the AI companies can argue their output constitutes transformative work or whether it’s an unapproved (illegal) derivative work of the original components.
It’s tough to know how these lawsuits will pan out, but what we do know is that, in some form or other, AI is with us to stay. Google is already scrambling to find ways to keep up with the new threat of tools like ChatGPT, and apps, websites, and even TikTok filters are quickly incorporating AI to make sure they’re staying ahead of the trend.
How will you use ChatGPT and AI in your daily lives?
Ben Tomlin
Ben combines his skills and passion for music and tech to provide photography and video/audio recordings for choirs and orchestras, as well as building new tools as a web developer.
Related Posts
February 4, 2023
How can musicians use AI today to improve their lives?
When it comes to capturing the essence and energy of a classical music concert,…
February 4, 2023
Top 10 Accounting Tips for Musicians
When it comes to capturing the essence and energy of a classical music concert,…