The new IPO Guidance is in two parts:

1. the legal framework for examining applications for or using AI and how this will be applied;
2. scenarios providing practical illustrations based on an abridged patent specification; each scenario includes a basic analysis showing how the IPO would apply the guidance.

In short, the guidance explains that:

• in the UK, patents are available for AI inventions in all fields of technology;
• AI inventions are typically computer-implemented and may rely on mathematical methods and computer programs in some way; UK patent law excludes from patent protection inventions relating solely to a mathematical method “as such” and/or a program for a computer “as such”; however, these exclusions are applied as a matter of “substance not form” by considering the task or process an AI invention performs when it runs;
• when the task or process performed by an AI invention reveals a technical contribution to the known art, the AI invention is not excluded and is patent-eligible;
• following HTC Europe Co Ltd v Apple Inc [2013] RPC 30, in which five “signposts” pointing to allowability were identified, an AI invention is likely to make a technical contribution if, when it runs on a computer, its instructions: (i) embody a technical process which exists outside the computer; or (ii) contribute to the solution of a technical problem lying outside the computer; or (iii) solve a technical problem lying within the computer itself; or (iv) define a new way of operating the computer in a technical sense;
• AI inventions are not excluded if they are claimed in hardware-only form, i.e. if they do not rely on program instructions or a programmable device for their implementation;
• an AI invention is only excluded from patent protection if it does not reveal a technical contribution; an AI invention is unlikely to make a technical contribution if its task or process: (i) relates solely to items listed as being excluded (e.g. a business method) and there is no more to it; or (ii) relates solely to processing or manipulating information or data and there is no more to it; or (iii) has the effect of just being a better or well-written program for a conventional computer and there is no more to it;
• these conditions apply whether the invention is categorised as “applied AI” or “core AI” or it relates to training an AI invention in some way;
• patent protection is available for training datasets when they are used in inventions which reveal a technical contribution; however, claims to datasets characterised solely by the information content of the dataset are likely excluded as presentation of information as such; and
• the sufficiency of disclosure of an AI invention or dataset is assessed, like any other invention, according to the principles set out in Eli Lilly v Human Genome Sciences [2008] RPC 2.

To access the new Guidance, click here.