In this week’s live Q&A on LinkedIn, the question came up when, where and how patentable AI innovations actually happen today. In this article, I’ll try to unpack the notion of AI patents from different angles, showcasing the immense impact of AI on the invention landscape today.
By the way, if you’ve missed the live Q&A on LinkedIn, I’ve put the recording on my podcast for you:
What are AI patents?
Whenever I encounter “AI & patents” in my day-to-day work, it’s in one of these three ways:
(i) People who invent patentable AIs: These are the brilliant human minds who develop groundbreaking AI technology, laying the foundation for transformative inventions across various industries. Patents in this category cover the AI as the protected invention, or as a building block in the invention.
(ii) AIs that invent patentable things: This is mostly about generative AI. This fascinating aspect involves AI systems generating novel inventions with limited human intervention, or maybe even completely autonomously. Here, the AI itself is the inventor, creating something new. The AI’s output is then patented.
(iii) Using AI in the patent profession: Although beyond the scope of this article, the intriguing potential of employing AI in the patent profession is a hot topic to be watched. Let me know if I should write more on this in another article.
For now, let’s focus on categories (i) and (ii) to understand how AI is revolutionizing the innovation landscape.
AI as a Building Block for Solving Technical Problems
This refers to category (i) mentioned above. AI’s pervasive influence in innovation is undeniable. As a patent attorney, I encounter cases every day where AI serves as a building block within the invention. AI is an enabling technology across industries, and is therefore running in virtually every digitalized product I can think of.
Here are some examples:
- AI for Predictive Maintenance in Factories: By analyzing vast streams of sensor data, AI can predict equipment failures and maintenance needs, optimizing manufacturing processes. A patent on this would cover the way how the AI produces its forecasts.
- AI for Circuit Layout Optimization: In the electrical control cabinet industry, AI algorithms can optimize circuit layouts, enhancing efficiency and minimizing errors. This is an example where, apart from the output of the AI; see category (ii), companies want to also patent the AI-based optimization tool.
- AI for Enhanced Image and Sound Quality: From cameras to microphones, AI is enhancing image and sound quality, enriching our multimedia experiences. Here, the AI would be part of the part of the patented camera, microphone, denoising method, and the like.
- AI for Smart Consumer Products: AI-powered electric toothbrushes adapt to individual user preferences, offering personalized oral care. Here, the AI would be part of the patented toothbrush.
- AI in Drug Discovery: In the biomedical field, AI assists in analyzing extensive datasets to identify potential new drugs and treatment strategies. This is again an example where both the AI’s output (the newly found drug) and the AI inside the screening tool is a target for patent protection.
And this is just the tip of the iceberg.
AI’s applications are virtually limitless, permeating every corner of our lives and industries. The companies behind these inventions strive to protect their novel implementations of AI through patents, ensuring they remain at the forefront of their respective domains.
AI as Inventor: The Unseen Creator
While category (i) highlights the ingenuity of human inventors behind AI, category (ii) brings forth a fascinating aspect often unnoticed by the public eye. AIs are quietly inventing new things all around us, contributing significantly to groundbreaking innovations.
Imagine an AI-driven model in the medical field that meticulously analyzes vast genetic data, ultimately leading to the discovery of a life-changing drug. In such cases, the AI acts as an inventor, at least as a co-inventor.
And my experience is that this happens all the time already today. It is just that many patent applications that cover AI-generated inventions don’t mention the involvement of AI in the invention process.
As European patent law primarily focuses on the novelty and inventiveness of an idea, the origin of the invention’s creation, whether human or AI, is not a determining factor. Thus, many patents resulting from AI-driven innovations don’t explicitly credit the AI for its contribution.
Naming an AI as an Inventor in a Patent Application
Of course, there is no chance to discuss this topic without touching on the question whether an AI can be officially designated as an inventor in a patent application. This question has been flagged to a broader audience by the famous “DABUS cases“. DABUS, an AI system, generated inventions, and patent applications were filed in various patent offices around the world naming DABUS as the sole inventor. But currently, most patent offices worldwide maintain the stance that inventors named in a patent application must be human beings.
Despite this, I personally don’t see a significant practical need to designate an AI as an inventor, and in fact none of my clients has ever asked me if that’s possible.
- When you’re reading an invention disclosure document, ask youself what the thing to be patented actually is:
- Should the patent cover a unique AI – to block your competitors from doing what your company does / from selling AI-equipped products?
- Should the patent cover the AI’s output – to block your competitors from selling the end product (be it created by AI or not)?
- Keep this in mind when you’re drafting the AI patent application, or when you’re reviewing your outside counsel’s draft.
Your next move
If you found this glimpse into the world of AI patents helpful, there’s more to come. As we move towards a future where AI is an integral part of invention and innovation, it’s crucial to stay informed about the latest developments. If you want to stay in the loop, sign up for my mailing list today. It comes with tons of added benefits.