Can flooding provide a safe harbour at the EPO?

Growing up, my parents’ house flooded on two occasions. I am therefore all too aware of the damage that floods can cause and the need to understand the risk of flooding occurring. As discussed in this article, assessing flood severity and risks is one application towards which AI is now being directed. I therefore thought the work discussed in the article would provide a good example for considering the patentability of AI at the EPO.

It is a common misunderstanding that computer programs (i.e. AI) cannot be protected using patents. Put simply, this is not necessarily the case. However, seeking patent protection for computer programs does have its own peculiarities and challenges that can be usually be overcome with professional expertise. For example, when applying for patent protection at the European Patent Office (EPO), an invention must produce a technical effect. For AI inventions (which the EPO consider to be mathematical methods) to have technical effect, they need to fall within one of two ‘safe harbours’ of being a technical application or technical implementation. A technical application would be a specific application of the AI to perform a technical task, such as image processing. A technical implementation would be a specific way in which the AI is implemented, such as the specific arrangement of hardware on which the AI is executed. If an AI invention cannot find a safe harbour, the EPO will likely consider the invention to lack technical effect and will refuse a corresponding patent application.

The EPO’s Guidelines set out more detail on when a mathematical method is considered to produce a technical effect. In particular, the Guidelines state that “if steps of a mathematical method are used to derive or predict the physical state of an existing real object from measurements of physical properties, as in the case of indirect measurements, those steps make a technical contribution regardless of what use is made of the results.” Flooding is a physical state of an existing real object and images provide indirect measurements. Accordingly, it follows that predicting flooding as discussed in the article should be recognised as a technical application of AI.

The same section of the Guidelines also provides a list of examples of applications that should be recognized as technical. One of these is “digital audio, image or video enhancement or analysis.” Generally, from experience, patent applications seeking broader protection are more likely to be considered allowable if directed to one of these technical applications. Therefore, as the AI described in the article uses image analysis there are additional reasons as to why it should be considered a technical application at the EPO.

To conclude, I would argue that using AI to assess flood severity or predicting flood risks based on real world image data would find safety in the first safe harbour of technical application. This is because the AI is being use for the specific application of predicting the state of a physical system, based on real world image data. If you’d like to discuss the patentability of your invention, whether it involves AI or not, please don’t hesitate to get in touch.