Legal reports, government consultations, and anime-style selfies rarely feel part of the same story – but in recent days, they have.
Tuesday, the U.S. Copyright Office released Part 2 In the long-awaited report on the copyright of AI-generated works.
What is its core information? Human creativity remains the basis of American copyright law, and AI-generated materials themselves are not qualified.
The office is clear. Individual tips, no matter how detailed or imaginative it is, are not enough. What is important is the author’s identity, and the author must involve human originality.
If a person curates, edits, or meaningfully converts AI output, that contribution can be protected. But the output of the machine itself? No.
In practice, this means that people who use text prompts to generate images may not have it in the traditional sense.
The report outlines three narrow schemes that may apply copyright: when assisted AI, perceived primitive human works merge, or when humans arrange the elements generated by AI in a creative way.
It sounds generous in some ways, but the fact remains that the court has been rejecting copyright claims for purely machine-made works, and the report affirms that position.
The Copyright Office is likened to a prompt to give photographers instructions: they may affect the results, but they do not improve the author’s level.
But just as re-drawing the line in Washington, Openai also urged British lawmakers to take a different path.
Wednesday, the company Submit a formal response Consult with the UK government’s AI and copyright.
Openai advocates “widespread text and data mining exceptions,” a legal framework that will allow AI developers to train publicly available data without first seeking permission from the rights holder.
The idea is to create a pro-innovation environment that supports investment and development in AI. In fact, let the machine read everything unless someone explicitly opts out. It is a position that has made Openai firmly conflict with many in the creative department, where the bells have been ringing for months.
Artists, authors and publishers regard the proposed exception as a backdoor license to scrape the network, turning years of human work into fuel for algorithmic engines.
Critics believe that even the exit model can put a burden on creators rather than companies and risk eroding the already fragile professional content economics.
Insert it into this copyright furnace is to release New research This week, the AI disclosed project’s claim this week that OpenAI’s latest model GPT-4O showed a suspicious understanding of paid content.
All of these take the more public (and very popular) example of AI blurring the boundaries: Studio Ghibli Trends.
Openai’s Image Generator made new improvements in Chatgpt last weekend, and although the studio’s co-founder publicly stated that he hated AI in 2016, it was all the rage for its ability to turn selfies into a Ghibli scene.
It’s so fun to see people like images in Chatgpt.
But our GPU is melting.
We will temporarily introduce certain rate limits while improving efficiency. Hope it won’t be long!
Free Chatgpt Free Layer will get 3 generations a day soon.
– Sam Ultraman (@sama) March 27, 2025
Career distillation into tips. Or is the creativity of AI really blooming in public consciousness?
None of this happened in isolation. Copyright law (historically slow and text-bound) is forced to change and adapt.
Governments, regulators, tech companies and creators are all scrambling to define rules or bend rules to make this debate better.
