How the EU AI Act will shape tomorrow's AI patents

On July 27^th, Elon Musk shared a deep fake video of US Vice President Kamala Harris on X, which received over 136 million views. In the video, she said things she never said.

Sparking widespread concern about the potential for AI-generated content to manipulate reality, its use has gained the attention of international law makers across the world, including the European Union (EU).

We already know that AI is leading the technological landscape with an 800% increase in AI inventions since 2017.

So, the EU has introduced the AI Act, the world's first comprehensive legal framework for AI to regulate its use and prevent malicious activities.

However, questions surrounding its patent eligibility, novelty, and inventive step are coming to the fore.

This article explores how the EU AI Act could reshape patent law, both in Europe and globally.

Understanding the EU AI Act

Coming into force on August 1^st, the EU AI Act aims to establish a unified regulatory framework for AI across the EU, addressing the risks and opportunities associated with AI technologies.  

To ensure the safety of the users, the parliament has made it a priority to make the AI systems more transparent, traceable, eco-friendly and non-discriminatory.

For that reason, rules have been set and a regulated, holistic framework for the fair use of Gen AI has been introduced. According to this framework, Europeans can trust what AI has to offer.  

This AI Act framework categorizes AI systems based on their perceived risk level, with high-risk AI systems subject to stricter requirements.

These requirements include transparency, accountability, and human oversight, ensuring that AI is developed and used in a responsible manner.

‍
Below is a high-level summary of the EU AI Act:  

Four-point summary

The AI Act classifies AI according to its risk:

• Unacceptable risks are prohibited (e.g. social rating systems and manipulative AI).
• Most of the text deals with high-risk AI systems that are regulated.
• A smaller part deals with limited-risk AI systems that have lower transparency requirements: Developers and operators must ensure that end users know that they are interacting with AI (chatbots and deepfakes).
• The lowest risk is unregulated (including most AI applications currently available on the EU single market, such as AI-powered video games and spam filters — at least in 2021; this is changing with generative AI).

Most obligations are made by providers (developers) of high-risk AI systems.

• Those who intend to market or start up high-risk AI systems in the EU, regardless of whether they are based in the EU or in a third country.
• And also providers from third countries who use the output of the high-risk AI system in the EU.

Users are natural or legal persons who use an AI system professionally, but not affected end users.

• Users (users) of high-risk AI systems have some obligations, although fewer than providers (developers).
• This applies to users in the EU and to users in third countries when the output of the AI system is used in the EU.

General purpose AI (GPAI):

• All providers of GPAI models must provide technical documentation and instructions for use, comply with the copyright directive, and publish a summary of the content used for training.
• Providers of GPAI models with free and open licenses only need to comply with copyright and publish the summary of training data, unless they pose a systemic risk.
• All providers of GPAI models that pose a systemic risk — whether open or closed — must also conduct model assessments and counter tests, track and report serious incidents, and ensure cybersecurity protection measures.

Patent eligibility and AI

Now you know what the EU AI Act implies, one of the most significant implications of the EU AI Act for patent law is its potential impact on patent eligibility.

Traditional criteria

Patent eligibility, historically, has relied on three main criteria:

1. Novelty: The invention must be new and not previously known.

2. Inventive step: The invention must not be obvious to a person skilled in the relevant field.

3. Industrial applicability: The invention must have practical use in an industry.

AI-specific challenges

AI-related inventions present unique challenges when it comes to meeting these criteria.

For example, the question of whether AI can be considered an inventor has stirred debate in legal circles.

In 2021, AI systems like "DABUS" filed patent applications listing the AI as the inventor.

However, it was later decided by the German Federal Patent Court as well as the UK Supreme Court that the AI system cannot be named as the inventor under the patent law.

While some jurisdictions, such as in South Africa, have granted patents to AI-generated inventions, many others, including the EU, have maintained that only human inventors can be listed on patent applications.

With these restrictions and paradoxes, AI-generated inventions raise questions about novelty and inventive steps.  

If an AI system can generate an invention that is truly novel and non-obvious, who should be considered the inventor?  

Should it be the AI system itself, the developer, or the owner of the AI system?  

The EU AI Act does not explicitly address these questions, but its emphasis on transparency and accountability may influence how patent offices interpret existing patent law.  

‍
Impact of the EU AI Act

‍
The EU AI Act’s emphasis on transparency and accountability may influence how patent eligibility is determined for AI-related inventions.

AI systems that contribute to the inventive process may be required to provide traceable and transparent outputs to demonstrate the human's role in the invention.

The Act’s human oversight provision suggests that inventions will need clear documentation of human involvement in the inventive process to be eligible for patents.

Novelty and inventive steps

AI-generated inventions

One of the most challenging aspects of patenting AI-generated inventions lies in determining their novelty and inventive step.

AI systems can generate innovations by analyzing vast datasets, creating solutions that a human inventor might never have conceived.

However, if an AI system is trained on existing knowledge, questions arise about whether its output is truly "novel" or simply a recombination of existing ideas.

Even if an AI system is used to generate an invention, the patent application may need to provide more information about the AI system's capabilities and how it was used to create the invention.  

This could make it more difficult to establish novelty and inventive step, particularly if the AI system's capabilities are not well understood.

Prior Art Considerations

The assessment of prior art—previous knowledge or inventions relevant to a patent claim—may become more complex in the context of AI.

As AI systems access and process massive amounts of data, there is a heightened possibility of inadvertently duplicating existing inventions.

The EU AI Act’s provisions on transparency and risk assessment could lead to more rigorous scrutiny of AI-generated patent claims to ensure that inventions are genuinely novel.

Additionally, the Act may push for AI systems used in innovation to disclose their sources of training data and methodologies.

This could help patent examiners assess whether an AI-generated invention meets the novelty requirement or if it is a derivative of pre-existing knowledge.

Possible changes in standards

The introduction of the EU AI Act could lead to a tightening of standards for AI-related patents.

The requirement for transparency and explainability in AI systems may translate into stricter scrutiny of AI-generated inventions.  

Patent applicants may need to provide detailed explanations of how their AI systems arrived at a novel or inventive solution.

Conversely, if AI's role in innovation continues to expand, there may be pressure to relax the standards, allowing for more leniency in determining novelty and inventive step for AI-related inventions.

Patent claim drafting and interpretation

Alignment with AI Act Principles

The principles of transparency and accountability enshrined in the EU AI Act may necessitate significant changes in how patent claims for AI-related inventions are drafted.

Claims will need to be written in a way that clearly explains the AI system's role in the invention process, ensuring that the invention complies with the Act’s requirements for human oversight and accountability.

For example, an AI-generated patent claim might need to specify how the AI system was designed, trained, and supervised to ensure the invention’s novelty and inventive step.

This level of detail will be essential to align patent claims with the principles of transparency outlined in the AI Act.

Possible changes in claim language

As AI systems become more integrated into the inventive process, patent claim language may need to evolve to reflect the unique characteristics of AI-generated inventions.

Current patent laws are largely based on human invention processes, but AI introduces new complexities.

For example, claims may need to explicitly address the role of machine learning algorithms, data sets, and neural networks in the invention.

Interpretation challenges

The intersection of the EU AI Act and patent law is likely to present interpretation challenges.

For instance, how should patent examiners interpret claims involving AI-generated inventions?

Should AI be treated as a tool, like any other machine, or does its autonomous nature require a different standard?

Additionally, how will courts handle disputes over patent claims that hinge on the workings of opaque AI systems, such as those driven by deep learning models that lack full explainability?

The EU AI Act may also influence the drafting and interpretation of patent claims related to AI-related inventions.

Patent claims must be clear, concise, and supported by the description in the patent application.  

In the context of AI, this may require patent applicants to provide more detailed information about the AI system's architecture, algorithms, and training data without compromising their trade secrets.

PATOffice and the low risk zone for AI

As a company that specializes in patent workflows, insights, and collaboration, the integration of Gen AI into PATOffice comes naturally.

We are in the low-risk zone of EU law, as the generative AI system was developed taking full compliance into account.

These help to simplify patent information and to interpret the findings in a completely transparent way.

The AI used by PATOffice supports without taking over the central decision-making processes and ensures that workflow inputs and judgments remain with human experts.

By complying with the EU AI Act, PATOffice responsibly integrates AI into its system without falling into the high-risk AI category, which would require additional regulatory review.

Conclusion

The EU AI Act has significant implications for patent law, particularly in the context of AI-related inventions.

While the Act does not explicitly address patent law, its requirements for transparency, accountability, and human oversight may influence how patent offices interpret existing patent law.  

As AI continues to evolve, it is likely that patent law will need to adapt to address the unique challenges and opportunities presented by this technology.

The EU AI Act represents a monumental step forward in regulating artificial intelligence, with significant implications for the future of patent law.  

Its provisions on transparency, accountability, and human oversight will likely reshape how patent eligibility, novelty, and inventive step are assessed for AI-related inventions.

Moving forward, patent law will need to adapt to the challenges posed by AI-generated innovations.

Policymakers and stakeholders must work together to strike a balance between encouraging innovation and ensuring the ethical and responsible use of AI.

The interplay between AI and patent law will continue to evolve, making it essential for legal frameworks to remain flexible and forward-thinking as AI systems become even more integrated into the invention process.

This ongoing dialogue between innovation and regulation will define the next era of technological progress, ensuring that AI contributes to society in a responsible and sustainable manner.