Introduction
Nowadays, we see that there is a fast-paced advancement in Artificial Intelligence globally, and eventually, this is a great thing for technological transformation. But the key issue is not technical at all. The fast-paced development in technology makes it difficult, from the legal perspective, to adapt to the simultaneous advancements. So, in this article, we will understand what the key legal developments in Artificial Intelligence and Intellectual Property are recently.
So, everything started with a painting, which was eventually a vibrant image—not created by human hands, but from the circuits of a machine. Stephen Thaler, who is the mind behind an AI system called the “Creativity Machine,” looked at the artwork it had created and thought: Why shouldn’t this be copyrighted? After all, it was completely original, creative, and undeniably different.
So, like every innovator might do, he also applied to register it with the U.S. Copyright Office. But the answer came swift and clear—a simple no. It was not because the work lacked creativity, but because it lacked the most basic thing: a human author.
And soon after this, the stage was set for Thaler V. Perlmutter, a case that would eventually echo far beyond one single image. The U.S. Court of Appeals for the D.C. Circuit ruled in 2025, making the message even louder: The court was of the view that copyright law, at least for now, belongs to humans. The court’s reasoning was that copyright requires human traits like intent, judgment, and the ability to understand the consequences of creation. Machines, no matter how intelligent they are, simply don’t fit under the same purview. But Thaler wasn’t the only one challenging the boundaries between human creativity and artificial intelligence.
When Ai Eats the Internet: Getty Images V. Stability AI Case
In order to understand this case, imagine a machine being trained on the entire internet, which includes millions of images, from cat memes to high-end stock photos. Now imagine that some of those images are coming from Getty, one of the largest image licensing companies in the world, and were used without any prior permission.
That’s exactly the accusation made by Getty Images against Stability AI for doing this with its model, Stable Diffusion. Getty Images filed lawsuits both in the U.S. and the UK. They claimed that millions of its images had been scraped and used to train AI models without any permission, which is a violation of copyrights, trademarks, and more.
In January 2025, the court allowed the suit to proceed, which now sets the stage for a blockbuster trial in June 2025. The outcome of this case will decide the future of AI companies—whether they need to pay for the data they use to train their models, or whether the Wild West of Artificial Intelligence data scraping will continue without any checks.
Songs in the Machine: Concord Music Group v. Anthropic
Now just imagine that you ask an AI chatbot to write the lyrics to a Taylor Swift song. Not something inspired by her work, but almost word for word. That is what Concord Music Group and other publishers have said Claude, an AI by Anthropic, can eventually do. They sued Anthropic for allegedly training the chatbot on millions of lyrics protected by copyright. In March 2025, a judge simply threw out some of their claims but left the door open for the most serious one: a direct copyright infringement case. Music, which is protected by tight copyright controls, may become the next major battlefield for AI in courtrooms.
The Code Wars: Developers vs. GitHub Copilot
This is not a recent development of 2025 but is still very relevant to understand the recent ones. Back in 2022, some software developers noticed something very strange. GitHub’s new AI tool, known as Copilot, could write several codes that were almost identical to what they had uploaded to their open-source platforms. Many of them were stunned by this and questioned whether their work had been used to train the AI without any credit or consent.
This further led to Doe Vs. GitHub, a case that tested whether Copilot had been violating the Digital Millennium Copyright Act (DMCA). But then, in early 2024, the judges said no; the outputs were not identical enough. The case has now moved to appeal, which may finally clarify how closely AI-generated content must resemble original content to be considered infringing.
A Comedian Takes the Stand: Silverman v. OpenAI
Recently, even comedians are getting serious about copyrights. Sarah Silverman and a group of other authors sued OpenAI in 2023, saying their books were used to train ChatGPT without any permission. In early 2025, the court dismissed several claims but left one standing: direct copyright infringement. The case later moved to New York, where it became part of larger consolidated lawsuits involving several authors. The authors have argued that AI is profiting by using their words. Now, this could be a landmark case, as it will determine how future books are written and by whom.
A Legal Research Revolution or rip off? Thomson Reuters v. Ross Intelligence
Recently, a legal tech startup known as Ross Intelligence came up with a big idea: to build an AI that can do legal research faster than Westlaw. But in order to execute this idea, Ross Intelligence was alleged to have copied thousands of Westlaw’s headnotes without any permission.
That did not sit well with Thomson Reuters, the legal giant behind Westlaw. And in February 2025, a U.S. District Court passed an order that actions by Ross were not within the radar of fair use, especially because the copying was commercial and harmed Westlaw’s business. This is the first major case to say: You can’t simply scrape legal databases and call it innovation.
ANI v. OpenAI: India Joins the Fray
Also joining them, the news agency Asian News International (ANI) has sued OpenAI. They have claimed that ChatGPT has been using it’s copyrighted content to train itself. The Delhi High Court responded seriously to the issue, issued summons to OpenAI, and appointed an amicus curia to advise the court. This case could eventually become a landmark ruling in AI and copyright. This is also a clear sign that AI regulation is no longer just a Western issue.
AI and Patents: The Machine-Learning Myth?
In Recentive Analytics v. Fox Corp, the federal court weighed in on another frontier: patents. Recentive had recently patented a machine-learning method applied to TV viewership data and sued Fox for infringement. But in April 2025, the court decided that you cannot patent generic AI methods applied to new fields unless you are actually improving the technology itself. The message is very clear: AI isn’t magic. It’s math, which is not patentable on its own.
How the European Union Is Changing the Game for AI and Copyright: The European Union AI Act(In Force 2025)
In year 2025, the European Union took a massive step, with the introduction of AI act: the first of its kind in the world. It doesn’t just regulate safety and transparency. The article 53 of the acts, in particular, zeroes in on something which is more sensitive about the use of copyrighted works in the training of AI models. EU is of the view. “If you want to train AI, you must know where your data is coming from” That’s actually a very big deal because when companied often uses massive datasets scraped from internet not always having knowledge what is inside it. But the catch is that: even though the law demands transparency, in order the trace the exact origin of every piece o training data is nearly impossible with today’s technology. That creates a legal tension between enforcement and its practicality. Still, this is a message for all tech giants to respect creators, or simply face the consequences.
The UK’s Careful Dance: Encouraging AI Without Leaving Creators Behind; UK Text and Data Mining (TDM) Consultation(Closes 25 Feb 2025)
The UK is eventually taking more measured approach across the channel. In early 2025, the government launched a public consultation in order to explore expanding text and data mining exceptions that is there even for the commercial use of AI. This is an attempt to keep it up with both EU and AI innovation. The idea is simple, let the developers train their models using publicly available contents but with a safeguard, and the rights holder have a discretion to opt out from the same. It’s considered as a tricky dance, while encouraging innovation without alienating the artists, publishers, and copyright owners. Simply, the UK is trying to find the sweet spot before the technology sprints too far ahead.
Key Understandings and Conclusion
These recent Developments on AI and Intellectual Property Laws from courtrooms in Washington, D.C., to New Delhi and London: one thing is very clear: the world’s legal systems are racing to keep up with AI. Each ruling, whether it is on music, images, code, or news will eventually shape the boundaries of creativity, ownership, and innovation in the age of AI. The major question is not just about who wants what but what actually it means to create in a world where machines can reproduce almost everything.
Author Name- Himanshu Yadav is Law Student, B.Com. LL.B. (Hons.) at Tamil Nadu National Law University
