AI, Copyright, and the Future of Artists Like Us
The U.S. Copyright Office just dropped a deep dive into AI and copyright. If you’re an artist, designer, or creative of any kind, you need to pay attention. The report, Identifying the Economic Implications of Artificial Intelligence for Copyright Policy, doesn’t just analyze the law—it breaks down how AI is reshaping the economics of creativity. And yeah, it’s as serious as it sounds.
Here’s what stood out:
1️⃣ Can AI Art Be Copyrighted?
Right now, AI-generated work isn’t eligible for copyright unless there’s enough human creativity involved. But where’s the line? If an AI “assists” your process, does that count? If AI spits out a design that’s 90% your style, should it be protected? After all, you can train AI on your style. These are questions no one has clear answers to yet, and that’s a problem.
2️⃣ AI’s "Inspiration" vs. Straight-Up Infringement
AI models are trained on billions of images, artworks, and written works—many created by human artists like us. But is that “learning” or theft? If AI generates something strikingly similar to an existing painting, is it just a remix, or is it copying? The way copyright law defines “originality” and “infringement” is about to get tested in ways we’ve never seen before.
3️⃣ Whose Face, Whose Voice?
AI isn’t just mimicking styles—it’s cloning identities. From digital actors to AI-generated voices, the tech is getting scary good at replacing human creatives. The report talks about rights of publicity—meaning, should artists, musicians, and performers have legal control over their likeness, style, or voice when AI can imitate them in seconds?
4️⃣ The Data Dilemma: Should AI Companies Pay for Training Data?
Right now, AI companies are scraping massive datasets—including copyrighted works—without paying creators. Some argue it’s fair use; others say it’s flat-out exploitation. If AI companies had to license our work to train their models, how would that change the game?
🔹 What This Means for Artists
We’re at a turning point. AI is already competing with us, and if we don’t push for fair policies now, the future could look like a digital Wild West where human artists are devalued while AI profits from our creativity. This report doesn’t provide final answers, but it does frame the economic stakes in a way policymakers can’t ignore.
💡 What’s Next?
Copyright policy is still evolving, and artists need to be part of the conversation. Whether it’s fighting for licensing rights, pushing for transparency in AI training, or figuring out how to protect our styles, voices, and livelihoods, we can’t afford to sit this one out.
👉🏽 What do you think? Should AI be able to train on copyrighted art? Should AI-generated work be copyrightable? Let’s talk. 👇🏽
text copyright 2025 Howard Simpson
Well of course AI should be able to use copyrighted works in its training. I can use copyrighted works in my training, can't I? My life experience, memory, personal imprint of-and-on my culture is derived from, and is emergent from, my experiences. While it wouldn't be useless, AI would be far less useful with the gaping hole in its training data defined as "all copyrighted works used without permission." There aren't that many AI companies, and for training, just like at home, I need one copy of the work. One. So in terms of harm, Anthropic, Open AI, Microsoft, Google, and a few others perhaps need to pay for that one copy? "How have you been harmed" is part of the damage calculation. "One publisher or artist saying "we didn't get our $350" isn't going to cut it. Yes, the total for all works used will be much higher, but it's *one* copy of each work. The argument hugely falls flat for the "re-use" where the AI might hypothetically "publish" the entire work for me to read, on demand. They don't do that. They aren't designed to do that. They're not stealing and re-publishing the work. Fair use or not, many artists and creators might be thrilled for thousands of viewers to be exposed to their work, and the corpus of human knowledge to be influenced by its inclusion. And to back up a bit, if it was "scraped"? Well then someone else already had it and likely(?) "published" the work without compensation to the author! It had to come from somewhere--someone got a copy somehow, and whether or not their wanton exposure of the work to the Internet at-large is for-profit or not, I'd rather not have a copyright holder bame ME for reading it if it shows up on my screen. Who am I? Author of multiple published-by-a-real-publisher books, musician with over a dozen albums published, and so on. I'm one of you. If I find out someone used one of my works, my first thought sure isn't "mechanicals," "royalties," or the like. It's "Cool! More people have seen my work. Not that many people are making a living off of writing, drawing, painting, performing, comparatively. That still makes it all worth doing. Just create!
ReplyDelete