Alphabet Soup
The moving target where AI, ML and IP converge.
I deal a decent bit with IP (intellectual property) in advertising in regards to artists' rights and licensing for things like imagery, music and talent. I try to follow along with what's happening with AI (artificial intelligence AKA machine learning) closely, because it is so tied to IP. What is cool, but what is ethical? This issue is dedicated not so much to answers since the landscape is changing quickly, but more about questions and how we can educate ourselves moving forward. There are lots of folks on the AI hype train right now. In a video (worth watching), Dr. Li Jiang, Director of Stanford AIRE program said it best, "So a job, once the machines can do it, let them do it, and then we focus on the human part."
Vivek Jayaram
He's an IP lawyer, and I really appreciate his bite-sized, yet informative takes on what is happening where IP and AI converge. Here, he explains why images generated from AI are not currently copyrightable. The criteria are that it must be 1. Original 2. Fixed to a tangible medium, and 3. Have human authorship. He explains:
Here's another useful take on why you might get yourself in trouble using prompts that include an artist's name.
Basic Licensing for Advertising
This is a hopefully helpful highlights of the most common licensing for advertising. I created this for internal use, and I am not a lawyer nor am I giving you legal advice. But if you've ever wondered exactly what Licensing means or what Royalty Free is, this is for you.
Book Club
In the education department, I came across this list of Best Books on Machine Learning, but I liked the way the interviewee, Eric Siegel, talked about machine learning so much, I'm going to pick up one of his books.
Machine learning is a real technology with a proven track record and a true value proposition, whereas AI is the brand we hear about. It’s a subjective word that people use in many different ways, and no one can agree on the definition. Generally, AI conveys at least a little bit of overpromising, that is to say: hype. It’s basically a brand on top of machine learning.
Last Thought
Hito Seryl writes,
"Visuals created by [machine learning] tools are statistical renderings… They converge around the average, the median; hallucinated mediocrity."
For the Designer
There are lots of cool things about AI, but you shouldn't feel pressured to use it in a way that is unethical for you. For me, I have some issues with the way the models are trained – both because of lack of permissions in the data used and because of bias in that data. I would also rather pay a human illustrator or photographer for their work. You have to decide how you feel about it. Some AI programs (Adobe is one example) use only licensed work to train their models and so have fewer data training concerns. Not all AI is about creating imagery – some are productivity helpers that may be really useful to you.
For the Client
If your Designer is hesitant to just go create an AI image for your project, they have good reason. There are tons of lawsuits happening right now because of lack of consent to use the data some AI models have trained on. They may not want to expose themselves to any legal issues and/or it may collide with their ethics. For perspective, AI is not actually some new technology. If you've ever used Siri or another digital assistant, you know it's been around a while. The difference is the access to the technology. And we're all just figuring out how to use it and what its real strengths are.