The link is massive – examiners and judges have to evaluate huge amounts of material
Bonadio 21 – Enrico Bonadio (Lecturer at City University London – The City Law School, PhD from University of Florence, and LLB from University of Pisa), Luke McDonagh (Assistant Professor of Law at London School of Economics, PhD from Queen Mary, University of London, and LLM from London School of Economics), and Plamen Dinev (Lecturer at Goldsmiths University of London, PhD in Intellectual Property Law at City Law School, and LLM at Leiden University). “Artificial Intelligence as Inventor: Exploring the Consequences for Patent Law” Intellectual Property Quarterly 1 (2021) 48-66. Accessed October 18, 2022. https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3798767
An overly expansive approach to obviousness would therefore mean that few things would be eligible for patent protection.69 Consider the example of combination patents. Whereas a human may not necessarily think of combining cooking recipes with recent advances in the medical sector, a computer would not be limited by such restrictions.70 Therefore, the more prior art that can be taken into account, the less likely an invention would be regarded as nonobvious.71 Moreover, while stricter patent procedures are to be generally welcomed (especially as patent offices are often criticised for being overly generous), a patent system where the benchmark involves a “skilled computer” is not without its problems. Examiners and judges would have to put themselves in the mindset of a machine and consider prior art which may be within the reach of computers, but not human beings. 72 This might be practically impossible.