Navigating the AI revolution in media while protecting & scaling Intellectual property

Charlotte Lund Thomson is a lawyer specializing in intellectual property (IP) and Policy. She mainly advises media companies and media professionals all over Europe and the US. Also joining was Alejandra Panighi, responsible for EU institutions' strategic relations at Mediapro. Mediapro is a multimedia communications group in Spain founded in 1994 and the company is well involved in movie and television production. And last but not least, Margarita Grubina, VP business growth at Respeecher. Respeecher was founded in 2018 in Ukraine and is a software company developing speech synthesis software enabling one person to speak in the voice of another particular person using AI. 

IP is the cornerstone of lasting businesses  

The European audiovisual sector has a big impact on the economy. It is a very entrepreneurial sector with over 130 000 businesses. 99% are SME’s that rely on IP as a foundation for their business model. They employ over 2 million people and contribute close to 47 billion euro.  

Intellectual property is at the foundation of these businesses because it promotes the originality and innovation of the company. This is what differentiates them from their competitors and enables them to monetize their services.  

For big production companies such as Mediapro, which exists for over 30 years and has gone through different technological changes -from physical products such as cassettes to the rise of internet, streaming services and now AI-, IP is the cornerstone of their business and they will do anything to protect it and keep it strong. That is why in this digital age, it’s important to be well informed and make the right strategic decisions when using AI. 

How to navigate the challenges as media companies  

There are lots of challenges with AI now, especially with generative AI, because it affects the value chain. But being aware of the fact that AI is transforming the sector is important to adapt and integrate new ways of monetization while preserving actual business models.  

Part of the solution is hiring experts and creating internal teams who can explore implementation opportunities and support other teams to find out where in the production process AI could bring value. Next to that, it’s important to test the technology available to see what it can do, even if you do not plan to immediately implement or commercialize it.  Mediapro for example uses AI to help in subtitling and has already implemented AI in different internal processes. They also experiment with genAI and voice cloning, not to commercialize it but mostly to fully understand how to work when so many fundamental rights and IP rights are involved in the process.  

From fear to opportunities  

Respeecher on the other hand found a way to monetise and scale IP, which is at the core of their business model. They decided to focus on ethical AI usage by applying 2 principles.  

Firstly, Respeecher will not clone voices without the permission of the owner or of their relatives in case the owner has passed. This is simply illegal. Second, it’s extremely important for Respeecher to not train their models on online available public data or on data that has been given by clients, because ownership is not clear in these cases. By applying these 2 principles they take away the biggest fear clients might have when cloning their voices: the fear of losing control over what is done with their voice IP.  

Because of those fears education has been part of Respeechers mission. It’s important to understand what is possible and what is not. For Respeecher the core technology is speech to speech, which means the voice of a real person is necessary to generate new content, so actors are actually essential and irreplaceable.  

One of the big opportunities with AI now is mass personalization. This year, Respeecher was able to generate 7 billion personalized messages from only one voice for a big sport event. What’s important to understand here is that although that sounds impressive, AI is not a magic button that you just push, and it does the work for you. It’s a complex process and requires skilled people and different trials to get it right.   

The key concept is approval. 

When stepping into a project that involves AI, it’s important to explain to creative contributors what you would like to do and get them to approve it in a formal way. You don’t want to be stopped in the production process by an unhappy contributor who didn’t understand or give approval.  

Next up is transparency and collaboration. Having clear communication during the process about what you will do with the materials after the project, who will be the owner and what will happen after the project is finished are important questions to answer.  Make sure that nobody continues training on your IP and that you own the rights to the output.  

As a producer, your IP is legally copyrighted due to the human creation and originality of the product. But when your content is 100% AI generated, it’s hard to obtain copyright protection, leading to difficulties in protecting the investment as well. The panel therefore advises you to keep a record of your gen AI production process. This way, you can prove human input and creativity if there is ever any discussion about the process.  

Europe regulating versus US innovating  

As Europeans, we sometimes complain about the number of rules and regulations. Is there a risk that the US has a competitive advantage due to more flexible AI laws? According to Charlotte Lund Thomson, the legal approaches are different. In Europe, we have very detailed legislation to avoid having to go to court, while the US is more case law based. US legislation is less detailed, and companies go to court more frequently to judge case by case.  Overall, the legal principles around copyright and personality rights are very similar in the US and Europe, but you indeed see a more risk-centered approach in the US.  

However, this risk-taking approach is best to be avoided. For example, if your expensive production is stopped during distribution because you didn’t receive the rights for voice cloning, you are not only liable to all other contributors but also liable to the distribution platforms. Lastly, it’s important to note that in the US legislation is more at state level and the laws in LA for example are very similar to Europe when it comes to AI. So, in the end the results are not so different, but we get to it with different legal approaches. 

Conclusions 

With IP being at the foundation of all media companies, it’s important to manage it correctly to guarantee the growth of your company. AI is changing the media landscape and it’s important to jump on that wave to stay competitive in this industry. Furthermore, AI is an amazing tool that brings a lot of new opportunities to scale IP and creates new business opportunities to monetize your assets.  

However, a good understanding of AI and working with ethical partners is essential. All parties need to have a clear agreement on how the technology will be used and implemented in the project and the collaboration and results need to be transparent to guarantee trust and avoid legal challenges. Finally, testing with AI can be beneficial for the growth of a company, it opens doors for new business model opportunities. Our European legal system is there to protect these opportunities. Yes, the laws are becoming specific when it comes to technology, but let’s use it to our own advantage to experiment safely.