The Great Intellectual Property Heist: Who Owns Your Creative Work Now?
The Great Intellectual Property Heist: Who Owns Your Creative Work Now?
The fundamental idea of ownership is breaking down. We are in the middle of a massive scrap over intellectual property (IP), and the stakes are enormous. It is no longer just a theoretical debate about who owns an AI-generated image. It is a genuine legal and commercial mess.
Hollywood is Making Peace, Not War
A massive piece of news this week shows one side of this fight. Actors Michael Caine and Matthew McConaughey have signed deals with AI company ElevenLabs. They are licensing their voices for AI cloning, with Caine suggesting it is about "amplifying" humanity, not replacing it. McConaughey, who is also an investor, will use his cloned voice to translate his newsletter into Spanish audio.
This approach is clever. Instead of banning the technology, they are putting a price on their specific IP: their unique voice. They are shifting the conversation from if AI will use their work to how much AI will pay for it.
The Studios Are Fighting Back
Meanwhile, the other major development shows the opposite end of the fight. The Content Overseas Distribution Association (CODA) in Japan, which represents various publishers, sent a letter to OpenAI asking them to stop training their AI models on copyrighted content. The group says that using their creative work without permission is a copyright violation under Japanese law.
This is not about stopping AI-generated knock-offs; it is about the huge data pool that AI models ingest. The studios are effectively saying, "You cannot eat our dinner to learn how to cook without paying for the ingredients." This fight is about the bedrock of the creative economy itself.
The Boss's IP vs. The Team's Agent
This whole discussion brings me back to a conversation I had last week. A contact of mine is building an internal AI assistant or agent, training it on their company's documents, processes, and their own deep expertise. The question he was asking was: who owns the resulting IP? Him, or the company?
My contact can obviously take the skills he learned with him. But what about the specific custom agent he built? The IP of the output (the code, the report) generally belongs to the company under "work-for-hire" rules. That part is usually sorted out in AI platform terms.
The real grey area is the knowledge itself. Your expertise, your know-how, the stuff you have banked over ten years in the job. When that is used to train an AI, is that now part of the company's asset base? Yes, mostly. You were employed to apply that knowledge, and the company has simply found a new, automated way to bottle it. Ownership of the actual agent will likely fall to the employer, since the tool was built using company resources and for company benefit. If he leaves, the company will likely retain access to that specific agent.
The Uncomfortable Truth
I use AI every day. That makes me personally part of the problem while I also appreciate the creative work that is being threatened. We all are. This is the messy reality.
The key takeaway is that the big players like Caine and the Japanese publishers are forcing a necessary re-evaluation of IP. They are defining the two paths forward: the one where you licence your genius for profit, and the one where you fight to protect the data that trained the machine. This fight is not about stopping progress; it is about making sure that the new AI economy pays its bills. The next few years will decide who is allowed to keep the fruits of the mind.
AI News This Week
We scan the headlines so you do not have to. Here are the most important stories and why they matter.
Amazon Goes After Perplexity's Shopping Bot
Amazon sent a legal threat to Perplexity over its Comet browser assistant. The e-commerce giant claims the AI agent violates its terms by accessing the platform without permission to shop for users. Perplexity countered that Comet is a user agent, not a scraper, and is only acting on the user's behalf. This signals a massive turf war as companies like Amazon, which has its own AI shopper called Rufus, fight to control the AI layer of online shopping.
Apple Pays Google 1 Billion for Siri Power
A new report says Apple is nearing a deal to pay Google around $1 billion annually for a custom Gemini AI model to power the next version of Siri. Apple will reportedly use this new 1.2 trillion parameter model for complex tasks like summarisation, but plans to keep Google's involvement quiet. This shows even the biggest tech companies realise building competitive foundation models from scratch is incredibly difficult and expensive.
AI Deals Bring Celebrity Voices to the Masses
ElevenLabs has landed deals with Oscar-winning actors Michael Caine and Matthew McConaughey to license their voices for AI cloning. Caine, the 92-year-old British actor, said the innovation is meant to "amplify" humanity, not replace it. The voice clones will be available through the company's Iconic Voices Marketplace, giving brands official, licensed access to use celebrity audio.
The Threat of Shadow AI is Growing
The unsanctioned use of generative AI tools like ChatGPT by employees without IT approval is a growing security blind spot known as Shadow AI. This creates risks of data leakage and non-compliance, with over a third of employees admitting to sharing sensitive work data with AI tools. With agentic AI and browser extensions now in the mix, the security risk goes beyond just chatbots.
Japanese Publishers Request OpenAI Stop Training
Japan's Content Overseas Distribution Association (CODA), representing various major publishers, has asked OpenAI to stop using their copyrighted content for training its models without permission. The request follows a rise in AI-generated images and videos that mimic the distinct styles of CODA members. This highlights the international and structural conflict over who pays for the data powering the AI boom.
Google's NotebookLM Adds New Features
Google's AI-powered research tool, NotebookLM, is adding deeper research features and support for more file types. This allows users to better interrogate source documents and notes, turning the tool into a more comprehensive AI research assistant.
Tool Review: Gamma
What it is
Gamma is an AI presentation maker that creates slide decks or documents from a simple text prompt.
What does it do?
The core of Gamma is pure speed. You can input a structure or a few paragraphs and watch the app build a visually decent presentation in seconds. You get the main points, text, and accompanying images all built in front of you. Creating a single AI presentation uses about 40 credits on their system, while smaller edits use about 10 credits.
Pros
Pace: The speed is genuinely fast. It cuts out the painful blank page problem by giving you a well-formatted first draft immediately.
Customisation: You can change themes, colours, and formatting easily.
Free Trial: The free plan gives you 400 credits to start, which is plenty for simple projects and getting to know the tool.
Cons
Collaboration Cost: If you have a big team, only paying users can co-edit decks.
Basic Themes: The pre-set themes are considered quite basic by some users.
Credit System: While paid plans offer unlimited AI creations, more advanced features and models cost credits (Plus gets 1,000 credits a month; Pro gets 4,000).
Use Cases
First Draft Decks: Perfect for quickly mocking up a structure for a client pitch or an internal update.
Content Repurposing: Turning a written report or a long email thread into a visual summary.
Low-Stakes Presentations: Building slides for a training session or a general information overview.
Verdict
I like Gamma. It does one thing, speed, and does it very well. I have the paid version because removing the Made by Gamma branding is worth the ten quid a month. But honestly, for most users, you can get a lot of value from the free version's 400 starter credits. If you spend too much time faffing about with alignment in PowerPoint, this is a genuine time-saver. It is not a full substitute for a design agency, but for getting from idea to first draft slides in a minute, it earns its money.