AI can’t be an inventor, but can transform the invention process. Explore how patent law addresses AI-assisted inventions in different jurisdictions.
Throughout history, patents have stood as symbols of human ingenuity. They serve as a way to reward creative minds that come up with solutions to technical problems.
But now, Artificial Intelligence (AI) is challenging the traditional notion of what it means to be an inventor. With AI-assisted inventions on the rise, patent offices and courts face a dilemma: How to deal with creations where humans are not the driving force?
In this blog post, we’ll explore how AI intersects with patent law, and whether a system built to recognize human inventiveness can accommodate non-human inventors.
DABUS: The first AI Inventor
In 2018, Dr Stephen Thaler stirred up a legal debate by filing “test” patent applications for inventions generated by DABUS, his AI system.
DABUS, or the “Device for the Autonomous Bootstrapping of Unified Sentience,” mimics human thought processes, enabling it to create new ideas autonomously. It came up with two inventions without a human becoming involved:
- A container for quickly heating food and drinks
- A light beacon for attracting attention in an emergency
The patent applications reported that the inventions were created by the AI of a machine named DABUS. Given that Dr Thaler was its owner, he claimed to have the right to the grant of the patents.
The USPTO, UKIPO, and EPO rejected his applications on the grounds that only humans can be inventors. South Africa, though, took a different stance and granted a patent listing DABUS as the inventor.
This sparked debate among IP experts, who argue that AI doesn’t fit the legal definition of an inventor under patent law. But the key point is that South Africa’s patent rules don’t include a specific definition of the inventor figure. This ambiguity leaves room for AI to be recognized as such.
So, the burning question remains: Can AI be considered a legitimate inventor? Let’s delve into USPTO’s perspective on this issue.
Main takeaways of the USPTO guidance
On February 13, the United States Patent and Trademark Office (USPTO) issued its Inventorship Guidance for AI-Assisted Inventions.
According to Kathi Vidal, Under Secretary of Commerce for Intellectual Property and Director of the USPTO, “The guidance strikes a balance between awarding patent protection to promote human ingenuity and investment for AI-assisted inventions while not unnecessarily locking up innovation for future developments”.
Here is a summary of the main points covered in the Guidance:
- Only natural persons can be listed as inventors on US patents and patent applications
In other words, AI systems can’t be named as inventors per se.
- To be acknowledged as an inventor when using an AI system, a person must “contribute significantly” to each claim in the patent application
The human contribution will be significant enough to qualify for a patent if it:
– Contributes significantly to the invention’s conception or reduction to practice.
– Makes a contribution that is meaningful when compared to the entire scope or scale of the invention.
– Does more than providing explanations of commonly known concepts or existing technologies to the actual inventors.
Other factors to consider when assessing inventorship for AI-assisted inventions are:
- Using an AI tool doesn’t diminish a person’s role as an inventor.
- Having a general idea or plan to solve a problem isn’t enough to be considered the inventor.
- Making an invention work isn’t the same as making a substantial contribution to its conception.
- Someone who develops a crucial component of the invention may be considered an inventor, even if they weren’t involved in every step of the process.
- Ownership of the AI system used doesn’t confer inventorship unless the person significantly contributed to the idea itself.
It’s important to note that this guidance is not yet a rule, as the USPTO is seeking public comments during a 90-day comment period.
How to go about claims covering AI inventions
Claims related to AI inventions should follow the same approach as other computer-implemented inventions. The description should provide detailed insights into the AI process, outlining how it enhances technical processes or resolves technical challenges.
AI inventions stand a higher chance of securing patent approval if claims focus on:
- How AI is architecturally integrated into a system.
- The sources of data used as input for an AI system.
- The features that set AI apart from other systems.
- Specific functionalities performed within an AI system.
Moreover, figures should show the hardware setup and include flowcharts depicting the AI process.
What lies ahead
The question of how the patent system should recognize AI’s role in inventions remains open. In this sense, a consultation conducted by the UK Intellectual Property Office in 2022 produced some interesting findings. It weighed options for reform, such as:
- Expanding the definition of ‘inventor’ to cover humans responsible for AI-generated inventions.
- Allowing AI to be recognized as the inventor.
- Finding a way to safeguard AI-devised inventions outside the patent system.
The UKIPO reported that most respondents preferred maintaining the current law status. Many voiced concerns about AI’s capability to invent without human input.
Now, patent law has consistently adapted to technological advances and inventors’ changing practices. And we are confident that courts and legislators will find a way to ensure that the patent system continues to foster innovation.
As we do at Montero Language Services! We stay on top of the latest technology trends: computer-Assisted Translation (CAT) tools, Machine Translation (MT), AI… But we don’t lose sight of the crucial role subject matter experts and designated project managers play in patent translation. Find out more about our solutions for the IP industry here.
