This image (AI, generated by Jack Galler) was the best-loved piece in the AI Turing test according to participants
A Turing test was performed with 11,000 people, participants were presented with two pieces of artwork, one human, and one AI generated, random chance would result in a score of 50%, and perfect skill 100%. The median score in this test was 60%, which is only slightly above random chance. This means that participants were only able to tell the difference between original, and AI art 10% better than if they had just guessed blindly. Even self-proclaimed “AI art haters” found themselves unknowingly preferring AI art over human art.
In another study, only 59.5% of respondents indicated that they understood what counted as Artificial Intelligence in the first place. If a little under half of people don’t even know what counts as AI, how in the world are they supposed to reliably tell the difference between original and AI art? The real answer is they can’t. The notion that art must have a ‘soul’ to hold meaning is being challenged more than ever., while art may hold meaning to some, that meaning is created by you, the beholder. In the same way that meaning can be applied to anything artificially or man-made.
With AI becoming better at managing prompts, and new technology being implemented to help non-artists bring their vision to reality without breaking the bank, commission art popularity could plummet significantly. The only art that AI cannot replace in the future is truly creative art that is very unique and difficult to come across. I would give Mapartche as an example for her effective use of emotion, or physical art such as murals, brush art, or graffiti artists. I’m sure there will be people who exclusively use and hire real artists over AI will exist but as time goes on their numbers will dwindle, and creatives will find other ways to get paid for their work. As we continue to advance AIs ability to imitate art, we also happen upon the gray legal consequences for artists.
There is a lot of uncertainty when it comes to the legal consequences of AI. To make AI art in the first place, the AI is trained off real artist’s work, many without their knowledge or consent. This initiates a lot of outcry from artists who believe their art is being stolen, and while intellectual property disputes are taking place, there is no clearly defined copyright law in place to protect this. While there are many proposed frameworks, there is no framework that is actually feasible. The problem is that it is impossible to quantify an artists “art-style” and therefore incredibly difficult or impossible to prosecute or create a reasonable legal framework for it in the first place. Because original works are used to train AI without the artist’s consent, courts will need to decide whether this critical fact infringes on “fair use” terms. Fair use terms generally depend on the “four factor test”; the purpose and character of the use, the nature of the original work, the amount and substantiality of the portion used, and the effect of the use on the market for the original work.
The authors of Rembrandt’s Missing Piece argue that AI generated art fails the four factor test for a few reasons. Commercial use generally weighs against fair use, and the content may not be considered transformative as the purpose isn’t different from the original content. Another important point is the fact that visual art is considered “highly creative”, and highly creative works are usually afforded stronger protection under copyright law. The most glaring argument made however is that AI art could directly compete with original works, which could be the nail in the coffin for commission artists as earlier mentioned. In short the pathway forward legally isn’t completely transparent yet.
AI art is advancing at a rapid pace, and will certainly not be vanishing anytime soon. Artists will have to adapt in new and innovative ways as time goes forward. The fate of AI art will also depend on how governments decide to deal the cards when it comes to copyright law, which will undoubtedly be a major force of change in regulation. If you still don’t believe me, take the AI art Turing test I mentioned at the beginning. I too thought I would be able to easily tell apart human artists from AI, yet it really puts into perspective how far AI has come in respect to authenticity. The results may surprise even those confident in their ability to differentiate human and AI art.
WORKS CITED
Alexander, Scott. “How Did You Do on the AI Art Turing Test?” How Did You Do On The AI Art Turing Test?, Astral Codex Ten, 20 Nov. 2024, www.astralcodexten.com/p/how-did-you-do-on-the-ai-art-turing.
Nader, K., Toprac, P., Scott, S. et al. Public understanding of artificial intelligence through entertainment media. AI & Soc 39, 713–726 (2024). https://doi.org/10.1007/s00146-022-01427-w
Pwinfield. “The Uncanny Threat: Unraveling the Dangers of AI in the Entertainment Industry.” Boesch Law Group, 29 Apr. 2024, www.boeschlawgroup.com/the-uncanny-threat-unraveling-the-dangers-of-ai-in-the-entertainment-industry/.
Pablo , et al. “ AI Royalties — an IP Framework to Compensate Artists & IP Holders for AI-Generated Content.” Cornell University, 5 Apr. 2024. https://doi.org/10.48550/arXiv.2406.11857
Polymenopoulou, Eleni. “Rembrandt’s missing piece: Ai Art and the fallacies of copyright law.” SSRN Electronic Journal, 2024, https://doi.org/10.2139/ssrn.4794932.
Leave a Reply
You must be logged in to post a comment.