Although artificial intelligence (AI) has been around for some time, the rapid development and widespread application of AI technology over the last 10 years has resulted in a global AI 'patent boom'.

This increase in patent activity has brought about questions around the patentability of AI inventions as well as debate as to whether patenting is the most suitable intellectual property (IP) protection strategy.

In the fourth instalment of the 'Tech Transfer and Innovation in the GCC' webinar series, our speakers discuss:

  • R&D and IP trends for emerging tech (AI, blockchain, FinTech, NFTs etc.)
  • How to identify and evaluate software-implemented inventions (with a focus on AI)
  • IP protection considerations for AI-related inventions
  • Patenting considerations for software and AI-related inventions
  • Protecting your data assets
  • The challenges in commercialising software and AI-related inventions
  • Takeaways for IP managers

Summary of takeaways

Identifying and evaluating software and AI inventions

  • Create a software-specific disclosure system. It's very common for developers to rely on open-source (OS) software during development to speed up software deployment and save on development costs. Using software disclosure forms, which ask general questions about the invention but also include specific questions about any open-source software used by the team, is a good way of capturing this information early on. OS software licenses, particularly commercial licenses, should be reviewed to ensure that any restrictions on patent ownership and commercialisation of derivative software are identified early and a strategy is put in place to work around those.
  • IP managers encourage researchers and developers to disclose their inventions to the IP manager (or responsible person) early on in the development process or when they have a Minimum Viable Product (MVP). This allows the IP manager to carry out an evaluation of the commercial viability and patentability of the software invention. A fully functional MVP decreases the commercialisation risk and allows them to test the performance of the invention and justify allocating more resources to develop and commercialise the technology.
  • Public disclosures are discouraged to avoid compromising patent rights until an evaluation is completed and a decision is made as to whether the technology should be patented.

IP protection considerations for software and AI inventions

  • All elements of an AI-related invention can be protected using a combination of the below:
    • IP rights (patents, copyright, designs)
    • Confidential information/trade secrets
    • Contracts (NDAs, IP assignments in agreements)
    • Internal systems and procedures to manage these assets
  • An IP strategy should be aligned with the business goals or commercial strategy for that invention. If secrecy is important, patenting may not be appropriate, as all patent applications are usually published. This also applies if the invention is easy to invent around or where it is difficult to detect patent infringement by third parties.

Patenting software and AI inventions

  • Software programmes or algorithms, as such, are not patentable. However, inventions that are implemented in software or involve AI may be patentable if they can meet the patentability requirements of novelty and inventiveness (or non-obviousness).
  • The same PTO guidelines that apply to computer-implemented inventions apply to AI inventions. The general rule is to frame the software/AI invention in terms of the problem it addresses (technical challenge) and how the problem is addressed (technical solution). Is the functioning of a computer another technology, or a technical field improved?
  • The debate around patenting AI centres around subject matter eligibility (given the software restrictions), sufficiency of disclosure (how much disclosure is required in a patent application) and inventorship of inventions made by AI (most countries agree that AI cannot be an inventor).

Protecting data assets

  • The widespread recognition of the value of data assets correlates with the increase in trade secret litigation.
  • In many cases, the data used to train an AI model or generated from the model is the most valuable asset. Data isn't patentable but can be protected in other ways, including as trade secrets and contracts.
  • Entities should consider instituting internal measures to identify, capture and track access to their trade secrets. During trade secret litigation, the threshold for showing 'reasonable steps' taken to protect trade secrets is high.
  • Measures include the use of NDAs, procedures for interviewing staff during on-boarding and off-boarding, follow through on confidentiality obligations on employees and third parties, data management systems and security systems to manage access.

Challenges in commercialisation

  • Creating a unique and innovative solution is only the first step in the commercialisation process. The challenges in bringing inventions to market include:
    • Having the right commercial strategy and business plan for the technology
    • Reaching mutual agreement with partners
    • Having the right start up team
    • Getting the technology to a mature stage of development to de-risk investment
    • Clarifying background and foreground IP
    • Identifying and reaching the right customers quickly
    • Access to funds
    • Educating investors on the legitimacy and value of other IP protection strategies (such as trade secrets)
  • These challenges generally exist for most types of inventions but are somewhat compounded when you are racing against time in a rapidly evolving tech area.

If you have any comments or would like to discuss further, please contact Tamara El-Shibib.

Panellists

Sean Flanigan
Director of Technology Transfer, KAUST, Saudi Arabia

Fawaz Al Qahtani
Acting Director of Industry Development and Knowledge Transfer, Development and Innovation, HBKU, Qatar Foundation, Qatar

Vivien Wei Cheng
Patent Attorney, JurisAsia LLC, Singapore

Matt Hervey
Partner, Head of AI, Gowling WLG, UK

Other webinars in this series