Can you patent Artificial Intelligence in the UK?

In September 2022 the UKIPO issued new, enhanced guidance to support those seeking a patent on AI inventions. This guidance falls out of a “call for views” on AI and IP 2021 and is based on an understanding of recent case law.

How do UKIPO define AI?

The UKIPO defines AI as “technologies with the ability to perform tasks that would otherwise require human intelligence, such as visual perception, speech recognition, and language translation.” AI inventions are considered to be computer-implemented and may rely on mathematical methods and computer programmes, which has been specifically excluded from patent protection by UK patent law.

Understanding when an AI invention be considered as eligible for patent

The published guidance is intended to provide a better understanding of how the law applies to AI inventions and assess patent eligibility, alongside examples and scenarios involving AI inventions.

In addition to the standard patent requirements of novelty, inventiveness, and industrial application, the UKIPO confirms in this guidance that patent law does indeed recognise AI as inventions, provided that there is a technical contribution to the state of the art.

When is an AI invention likely to make a technical contribution?

The guidance states that:

An AI invention is likely to make a technical contribution if, when it runs on a computer, its instructions are either:

• embody a technical process which exists outside the computer, or

• contribute to the solution of a technical problem lying outside the computer, or

• solve a technical problem lying within the computer itself, or

• define a new way of operating the computer in a technical sense

An AI invention is unlikely to make a technical contribution if its task or process either:

• relates solely to items listed as being excluded (e.g. a business method) and there is no more to it, or

• relates solely to processing or manipulating information or data and there is no more to it or

• has the effect of just being a better or well-written program for a conventional computer and there is no more to it

What is the difference between Applied AI and Core AI?

The UKIPO categorises AI in these guidelines as either Applied AI and Core AI. This categorisation of AI provides guidance to UKIPO examiners when assessing AI-related patent inventions, as well as providing guidance to applicants and inventors on what may be deemed patentable.

Applied AI: The UKIPO suggests that Applied AI inventions are those where the AI is used to contribute to the solution of a technical problem or perform a technical process in other fields other than AI or makes the computer run better. The patentability of applied AI related inventions is associated to the end use of the AI. It is in these cases where the AI invention is likely to reveal a technical contribution and may be eligible for patent protection.

Core AI: The UKIPO suggests that Core AI inventions are those where it advances in the field of AI itself such as algorithms or mode, improving the internal workings of a computer and revealing a technical contribution as such, rather than any real-world application of those models and algorithms.

Which AI inventions are likely to be patentable?

The UKIPO lists useful examples and scenarios of AI inventions that may be patentable.

Examples of Applied AI – where the AI is used to contribute to the solution of a technical problem or perform a technical process in other fields other than AI:

• Automatic number plate recognition system for recognising a vehicle registration

• Monitoring a gas supply system for faults

• Analysing and classifying movement from motion sensor data

• Detecting cavitation in a pumping system

• Controlling a fuel injector in a combustion engine

• Measuring percentage of blood leaving a heart

Let’s take the example of measuring percentage of blood leaving a heart. This is the AI invention. This would lead to an improved measurement of the blood leaving the heart and serves as the AI invention performing a technical process in other fields other than AI (Applied AI). This reveals a technical contribution of the AI invention and can be eligible for patent protection.

Examples of Applied AI – which makes computers run better

• Examples of Applied AI – which makes computers run better

• Cache management;

• Continuous user authentication;

• Virtual keyboard with predictive text entry

Examples of Core AI – advances in the field of AI itself such as algorithms or models

• Processing a neural network on a heterogeneous computing platform

• Special purpose processing unit for machine learning computations

• A multiprocessor topology adapted for machine learning

Which AI inventions are likely to be refused a patent?

The UKIPO lists useful examples and scenarios of AI inventions that are likely to be refused.

These are inventions that do not solve a technical problem within the computer or cause it to operate in a new technical sense.

Examples of Applied AI

• Automated financial instrument trading

• Analysing patient health records

• Identifying junk e-mail using a trained AI classifier

Examples of Core AI

• Optimising a neural network

• Avoiding unnecessary processing using a neural network

• Active training of a neural network

Transparency and certainty when applying for patents in AI

Whilst the topic of what AI-related inventions may be patented remains contested, these scenarios, alongside the guidelines outlined, provide helpful guidance in what applications may be considered eligible for patent protection and what may be refused.

The published guidance reaffirms what we already know; the UKIPO guidelines treat AI inventions and non-AI software inventions similarly in that they must satisfy the same patent eligibility criteria and must provide a ‘technical effect/contribution’.

Although AI-related inventions are examined on a case-by-case basis, this guidance provides increased transparency and certainty for inventors and applicants allowing them to compare their own inventions with these scenarios of Applied AI vs. Core AI.

It is also useful to UKIPO examiners when assessing AI-related inventions, furthering their understanding of what can be considered a technical contribution.

Overall, these guidelines demonstrate the opportunities for protection and are a good indication of the UKIPO’s effort to recognise AI-related inventions, moving more in line with the UK Government’s National AI Strategy published in 2021.

Source Advisors offers a unique IP harvesting service to help qualify patent ideas. Get in touch today.