It is a requirement of most, if not all, patent systems that an invention should be non-obvious over the state of the art in order to warrant patent protection. The test for an inventive step is intended to prevent the patenting of minor, routine improvements and incentivise more fundamental research. Inventive step is therefore a measure of the extent to which an invention goes beyond or above the state of the art. Strictly speaking, it is not a test of whether human ingenuity was actually involved in creating the invention. This is clear from the fact that accidental discoveries are patentable. Instead, the test asks a hypothetical question: would the invention have been obvious to the skilled person at the time the patent application was filed?
But what if the invention was the brainchild not of a skilled person but of a machine? The fourth industrial revolution looks set to change the face of science, providing researchers with access to new technologies that assist in the development of inventions. One example is the pharmaceutical industry, where AI is already being embraced for drug discovery. Companies are hunting for immune-oncology and cancer drugs and metabolic disease therapies using AI to identify patterns hidden in large volumes of data. Increased automation will also facilitate the large-scale screening of potential drug candidates identified by AI and the testing of known drugs for new uses. Combining increased automation with AI will create an even more powerful technique for identifying and developing potential new treatments.
Of course, AI has myriad potential applications outside of the pharmaceutical industry. AI has been used to design lighter and better parts for concept planes and improved car chassis based on the analysis of huge quantities of data. Companies such as Iprova use AI to help their customers create inventions based on points of convergence between previously unrelated industries. Their technology uses vast amounts of publicly available information, bringing together topics from seemingly distant areas and connecting them. Examples include the connection of a specific drug delivery technique to a high value oil exploration problem, and the use of advanced sensors built into an autonomous vehicle to take measurements from its passengers for healthcare applications.
So what happens when automated screening makes the testing of an osteoporosis drug for the treatment of breast cancer all but inevitable? Or when AI makes searching for the solution to an oil exploration problem in the field of drug discovery a matter of routine? Is the result obvious, even though it might surprise the leading human experts in that field? As the fourth industrial revolution takes hold, should the assessment of inventive step change?
As mentioned above, inventive step is considered from the point of view of the skilled person. In the UK and Europe, the skilled person is someone who is aware of what was common general knowledge at the time and who had access to everything in the state of the art, but who possesses no inventive ability themselves. Currently, the skilled person would not normally be expected to look for a solution to a technical problem in a technical field unrelated to that in which they are an expert. So, for example, a person skilled in oil exploration would not be expected to look for a solution to a problem connected with oil exploration in the field of drug discovery. Therefore, the application of known drug discovery technology in the field of oil exploration may well be considered inventive and therefore patentable. However, as discussed above, AI is now being used to make precisely this type of connection. In such cases, the inventive step of such inventions may result solely from the work of AI.
Should the definition of the skilled person now include those with access to AI capable of making this type of connection in unrelated technical fields? Similarly, in the face increasing automation in the field of drug discovery, should the definition of the skilled person eventually include those with access to technology allowing the large scale, automated screening of known drugs for new uses? Should the skilled person include the machines themselves?
In Europe, the skilled person is considered to have at his disposal the means and capacity for routine work and experimentation. If the use of AI and automated screening becomes routine, it is arguable that the definition of the skilled person should include those with access to these technologies, making it much less likely that the results are deemed to be inventive. However, this could have serious implications which undermine the purpose for which the patent system was developed.
Patents encourage innovation by giving the inventor the chance to recoup the investments made during the development process. It is arguable that, without patents, R&D spending would be significantly less or (at least in some fields) eliminated altogether, as third parties would be free to exploit any inventions. It is already the case that pharmaceutical companies frequently abandon promising drug candidates because of perceived weaknesses in their patent protection; understandable given the estimated US $2.6 billion price-tag for developing a new treatment. The patent system also encourages the disclosure of those new inventions which, if kept secret, would impede innovation.
Even in the face of increased automation and the use of AI, the development of new technologies will still be an expensive and time-consuming process in many fields, at least for the foreseeable future. Pharmaceuticals will still require years of expensive research even after potential treatments have been identified. New inventions arising from AI will still require development and testing. Furthermore, AI will only be able to build on what is known if it is able to access the necessary information. Encouraging innovators to divulge the details of their inventions will still be key to developing new ones.
If the definition of the skilled person is modified to reflect the new technologies arising out of the fourth industrial revolution, the bar for inventive step may be raised so high that the patent system is no longer an effective incentive to innovate. This raised bar would apply to all inventions, irrespective of whether AI and automated screening were actually used to arrive at the invention. Only those inventions which go above and beyond what AI and automation are capable of producing would be patentable, a requirement which may be harder and harder to meet. Inventive step should perhaps remain a measure of an invention’s contribution to the sum of human knowledge, irrespective of whether these new technologies exist to help us invent. Perhaps we will have to depart entirely from the anthropocentric “skilled-person” assessment of inventive step to avoid undermining the patent system’s crucial role in incentivising innovation.
If you have any questions or require further information, please do not hesitate to contact
Julia Venner, or your usual Kilburn & Strode advisor.