A new paradigm will shake up the IP landscape, as cognitive activities once performed by humans will now increasingly be performed by machines. Artificial intelligence and machine learning (collectively AI) are hot topics in almost every industry, affecting everything from robotics, autonomous vehicles, and consumer devices to health and pharmaceutical technologies. AI is being used to solve complex problems and improve decision making, as well as to develop new products and processes. AI systems utilize algorithms that enable those systems to learn and develop through the analysis of information they are provided, oftentimes without human intervention or instruction.

Who is the Inventor?

U.S. Patent Law defines the inventor as “the individual or, if a joint invention, the individuals collectively who invented or discovered the subject matter of the invention.” Similarly, Congress has stated that the Patent Act is intended to “include anything under the sun that is made by man”.

As existing law and court decisions appear to rule out a machine or program being an inventor, the question of ownership of AI inventions could leave many contributors to the development of an AI system tussling for rights. Let’s consider an example.

Company A develops an AI program or machine, which it sells to Company B. Company B operates that AI on resources owned by Company C, such as servers in a cloud computing environment. Company B also obtains data from Company D that is used to train the AI. After training, the AI produces an invention – so who is the inventor?

To qualify as at least a joint inventor, one must contribute to the conception of the claimed invention in a way that is not insignificant in quality, when the contribution is measured against the dimension of the full invention. Arguments could be made that at least Company B, that owns the AI program/machine, and Company D, that provided the training material, contributed significantly to the development of the AI. In particular, was it a group of individuals at Company B that selected data from Company D to train the AI and then “discovered the subject matter of the invention” produced by the AI in a way to qualify as inventors?

Barring contractual considerations, Companies A and C might also claim to be inventors because they produced or operated the AI that created the invention. However, their contributions might also be dismissed as being merely that of the toolmaker and the technician as opposed to that of the inventor that selected the AI tool and directed the technician’s application of ordinary skill.

Let´s also consider the impact if the training data is taken from the Internet or social media. If the AI scrapes a personal website and that data leads to the invention, the owner or provider of the site content might also claim inventorship.

AI as an Infringer.

Stephen Hawking once said: “The short-term impact of AI depends on who controls it; the long-term impact depends on whether it can be controlled at all.”

As with inventorship, existing laws and precedent do not specifically provide for a machine or program as an infringing party. Under U.S. federal law, infringement arises when “whoever without authority makes, offers to sell, or sells any patented invention.” As an AI system learns and modifies its external behaviors, it is possible that a resulting product or process might infringe one or more patent claims. To illustrate this issue, let’s consider an extension to the above example:

The AI program/machine developed by Company A, owned by Company B, operated on resources owned by Company C, and trained with data from Company D develops a product or process that infringes an issued patent -who is the infringer?

Company B, the owner of the AI program/machine, might be the most obvious target regardless of whether the company had any knowledge of the infringing activity. Ignoring contractual arrangements, an argument could be made that Company C, which is operating the AI and thus, arguably, directly or indirectly making the infringing product or practicing the infringing process, might also be considered to be an infringer.

An interesting wrinkle involves potential accusations of induced infringement. Under current law, the induced infringer is “whoever actively induces infringement of a patent.” The Federal Circuit has interpreted this to mean the alleged inducer must have knowingly aided another’s direct infringement of a patent.

An argument could be made that Company D, which provided the training data, induced infringement if the company knowingly aided the AI infringe the patent by providing particular training data that would lead to infringement. A similar argument could be made that Company A induced infringement if Company A instructed Company B’s use of the AI in a way that Company A knowingly aided the AI infringe the patent. If Company C which is operating the AI, is deemed the direct infringer, then Company B, which owns the AI and directed its operation, may be responsible for inducing the infringement.

Impact on Businesses

For now, there are many uncertainties in this new frontier and IP lawyers will need to work closely with patent holders and technologists to develop strategies for protecting assets and limiting potential liabilities within this space.

As part of a comprehensive review, the following issues may be relevant:

  • Are new agreements needed to ensure that inventions resulting from AI are owned by a specific entity?
  • Should joint development agreements be modified to ensure ownership of technology developed by AI?
  • Should AI’s access to specific data be restricted?
  • Are new indemnification agreements needed?
  • If user data will be used to train the AI, do user agreements or license agreements need to be modified?