Intellectual Property and Artificial Intelligence – Emphasis on Copyright

Decoding AI & IP: The Interaction Shaping Our Future

Artificial intelligence (AI) and intellectual property (IP) are no longer the domain of tech enthusiasts or legal experts. From the sidelines to center stage, these topics have taken the top in the global conversation. The more sophisticated AI is, the more complex its relationship with IP will be, raising questions that resonate far beyond the boardrooms of Silicon Valley or policy halls of Brussels.

To understand this dynamic, we must first begin with AI. At its core, AI is a product of the digital revolution, born from the simple computer. And, what started as a computer has now grown into a powerful force that is redefining industries, culture, and the way we think about innovation.

What Exactly Is AI?

To understand its scope, it helps to look at the three main types of AI based on capability: Narrow AI, General AI, and Super AI.

1. Narrow AI:The Specialist 

Narrow AI, often called 'Weak AI', is the workhorse of the AI world. It excels at specific tasks but lacks the ability to generalise. Think of Siri setting reminders, IBM Watson analysing medical data, or even ChatGPT generating text. These systems are powerful within their domains but cannot step outside of them.

While Narrow AI has revolutionised sectors like healthcare, finance, and customer service, its limitations are clear. It’s a tool, not a thinker. Yet, its success underscores how far we have come in harnessing AI for targeted applications.

2. Artificial General Intelligence (AGI): The Aspiration

AGI, or 'Strong AI', represents the next frontier: machines that can perform any intellectual task a human can, or even above human levels across a wide range. This is the stuff of science fiction – think robots that can write novels, solve complex problems, or even feel emotions.

But AGI remains elusive still. However, its development raises profound questions about ethics, autonomy, and the nature of intelligence itself. For instance, what happens when machines can think and reason like us – or outperform us?

3. Artificial Superintelligence (Super AI): The Hypothetical

This brings us to Artificial Superintelligence or 'Super AI': a theoretical idea where machines 'outsmart' human intelligence in every possible way. Imagine a system that not only solves problems faster than us, but also outdoes us creatively and even emotionally. It is a concept that is both exciting and disconcerting. Would or could such a futuristic system become a serious threat to humanity at large?

AI's Real-World Impact

Today, AI is no longer a sci-fi concept. Its ability to learn, adapt, and mimic aspects of human thought is increasing exponentially.

AI is changing the world, from healthcare to finance, but its impact on creative fields and IP is what appears to be generating a lot of buzz and plenty of legal headaches right now.

The European Union (EU), often seen as a leader for the regulation of technologies, recently passed the world's first legal framework relating to AI, namely the EU AI Act of 2024. This groundbreaking piece of legislation addresses issues like transparency, safety, and accountability.

The EU AI Act classifies the risks of AI into four levels: unacceptable, high, limited, and minimal. They apply to all AI systems that affect individuals in the EU or are offered on its market.

For example, developers of large language models are now required to disclose the source of their training data, emphasising transparency – a rule seen as key to fighting deepfakes. As a result, AI-generated output is to be clearly marked therein.

However, when it comes to AI-generated content, it is unclear who owns it or if AI infringes on rights that are already owned by others. The EU AI Act does not provide answers to this.

The EU AI Act itself does not contain explicit provisions dealing with IP rights, but it does require that providers of AI models abide by the obligations set forth in the EU's Copyright Directive (Directive 2019/790). This is particularly important when it comes to the scraping of databases, a crucial step in training AI systems.

A Global Patchwork

In general, in the EU, the copyright duration for a literary, dramatic, musical, or artistic work where the author cannot be identified, is 70 years from the date of creation. This is specified under the Term Directive (2006/116/EC). This directive provides a framework for copyright protection in cases where human authorship cannot be established, but it is not understood how it applies to works made by AI.

This gap in legislation becomes even more clear when compared to other regions. In the UK, for instance, copyright law protects computer-generated works for 50 years. However, following a recent public consultation, the UK government chose not to amend this rule opting instead for a wait-and-see approach while monitoring AI’s impact on creativity and IP.

Across the globe, China is charting its own course. While eager to position itself as a leader in AI innovation, it also prioritises a tight regulatory oversight. In July 2023, China introduced the Interim Measures for the Management of Generative Artificial Intelligence Services, requiring that such technologies respect IP rights and uphold commercial ethics. The move reflects China’s dual focus on fostering AI growth while mitigating potential risks.

However, an interesting judgement of Li v Liu delivered by the Beijing Internet Court recently granted copyright protection for a creation prompted through an AI tool, signalling a bold, if controversial, interpretation of IP law.

But there’s an even larger issue looming: the looming scarcity of data. Experts warn that as AI continues to advance, the availability of 'free' training data is dwindling. High-quality AI systems depend on enormous volumes of information, much of which is protected by IP laws. This raises a difficult question: how can we fuel innovation without infringing on the rights of creators?

Japan, for example, a major economic power, is prioritising innovation, potentially at the expense of creators' rights. By contrast, the EU is taking a cautious, regulation-heavy approach, while China is strategically positioning itself as a leader by balancing growth with IP protections.

It’s a patchwork of policies, and there’s no precise consensus. Could Japan’s full-speed-ahead path inspire others? Will China’s legal frameworks set new standards? And how will the EU’s deliberate pace fare in an era of rapid technological advancement?

The Core Dilemma: Authorship And Ownership

At the heart of these debates is a fundamental question: Who owns AI-generated output? Is it the programmer, the end user, or the AI itself? As data becomes the lifeblood of innovation, the race to control it intensifies.

This is not just talk. For artists, musicians, scientists, and businesses, this matters a lot, this is their life's work at stake. How can we push new ideas but safeguard IP rights simultaneously?

The Road Ahead

The conversation around AI and IP is just beginning. The question is not whether we shall find a way forward – but who and which plan will lead the way. As always, the future waits for no one. Are we ready, though?