Artificial intelligence (AI) has been identified as having many uses. One discussed recently, in a report commissioned by GCHQ, is how AI may be used to help with national security.
Artificial intelligence is good at spotting patterns in data that could be overlooked by human eyes and that require very large data sets to be analysed. For example, in the sphere of national security, AI may be able to assist with detecting and stopping cyber-attacks, which themselves may be being made using AI software.
However, the report argues that AI may find less relevance in trying to predict when terrorism offences may be committed. Is this because we have yet to develop true artificial intelligence and thus the systems humans build are only able to perform tasks that we train them to do?
It seems that the future may still be bright for human spies! However, AI might be there to assist them in analysing large data sets that are collected by intelligence agencies.
As I see it, from the intellectual property perspective, it is interesting to consider whether such AI innovations are likely to be patentable. Computational models and algorithms behind AI and machine learning may be of an abstract mathematical nature and thus, in at least some jurisdictions, may be prohibited from being granted patent protection.
However, when these models and algorithms are implemented in specific technologies, such that they contribute to providing a “technical effect” outside of the models and algorithms themselves, patent protection may be able to be granted for these kind of inventions.
As lots of uses for AI in national security may involve processing large sets of non-technical data, e.g. relating to the details and actions of individuals, it may be more difficult to obtains patents for these innovations. Furthermore, there is the consideration that national security agencies often would not look to patent their inventions, instead wanting to keep them secret (as part of the patent process, the patent application is published).