26 September 2022

Guidance on examining patent applications relating to artificial intelligence inventions in the UK

The UK Intellectual Property Office (UKIPO) released a guidance note for the examination of patent applications relating to artificial intelligence (AI) inventions. The UKIPO has confirmed that patents can be granted for AI inventions, given they provide a technical contribution to the state of the art.

Following a period of consultation that ran from 7 September to 30 November 2020, the new guidance note details the requirements of AI technologies to meet patentability criteria. As computer programs are specifically excluded from patentability criteria, the new guidance note and accompanying scenarios provide clarity when seeking to patent AI-based technologies.

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”.

In summary, the new guidance note states that:

  • AI patents are available for all fields of technology.
  • Whilst mathematical methods or computer programs are excluded from patent protection, when the task or process performed by an AI invention contains a technical contribution, it is not excluded.
  • An AI invention is likely to provide a technical contribution if it:
    • carries out or controls a technical process existing outside the computer;
    • contributes to the solution of a technical problem, external to the computer;
    • solves a technical problem in the computer itself; or
    • defines a new way of technically operating a computer.
  • AI inventions are not excluded if they are claimed in hardware-only form (ie. they don’t rely on program instructions or a programmable device).
  • AI inventions are likely to be excluded from patentability if they relate to an excluded item, relate solely to processing data, or are a general improvement on a program or conventional computer.

When considering the patentability of AI technology, focus is therefore placed on the technical contribution the invention makes to the state of the art.

Example scenarios

The UKIPO has also released a series of scenarios concerning AI or machine learning (ML) technologies and whether they meet the criteria for patentability. These scenarios focus on the issue of excluded matter and cover a breath of fields and technologies with worked examples of why each invention is or isn’t excluded from patentability.

Particularly interesting scenarios include the training of a neural network, in which the end result of the process and its intended use can be the deciding factor in whether or not it is excluded from patentability.

For example, training a neural network classifier system to detect cavitation in a pump system is allowed. Such a method involves correlating data pairs with class values to produce a training dataset (wherein each class value is indicative of an extent of cavitation within the pump system) and then training the neural network classifier system, using the training dataset and back propagation.

The fact that this process is reliant on a computer program does not exclude it from patentability, since it provides a contribution which uses physical data to train a classifier for a technical purpose – namely, the detection of cavitation in a pump system. The end result of this training, and its contribution, is therefore technical in nature.

By contrast, active training of a neural network is not allowed. Such a process involves determining areas of weakness in the neural network by comparing confidence levels to a threshold, then augmenting the training data with data related to the area of weakness. For example, a neural network used for detecting animals in pictures may struggle to identify cats, so the specimen data may be augmented with additional pictures of cats. This is more efficient than simply expanding the dataset across all elements.

While this method may result in a more efficient training method for a neural network, it does not itself produce a neural network that operates itself more effectively or efficiently. The mere identification of specific additional training data cannot be said to relate to a technical problem. As such, no technical problem has been solved within the neural network, and no technical effect is produced. A claim directed to this would therefore be excluded as a program for a computer as such.

Additional scenarios may be accessed here, and the guidelines published by the UKIPO are available here.