AI and Patents
We bring to Coevolve IT the experience of innovation, having converted ideas into descriptions that were used by patent attorneys to file patents that could be granted.
Training AI on text
Today AIs using Large Language Models (LLMs) raise questions about intellectual property because such AI tools have been trained on large bodies of text that the companies training them may not have had rights to.
Innovation required for patents
Does this affect patents? If your innovative concept is sufficiently different to be granted as a patent, maybe not, because your conceptual leap cannot be anticipated in training an AI or supplying prompts to it.
Patent examination
Before a patent is granted, a patent examiner will study it in relation to all prior patents arising from patent searches. This will include the use of resources such as those provided by the World Intellectual Property Organisation (WIPO). (See the slogan of the WIPO below.) Given that this is a manual, technical, search of specialised text, automation is possible. This could either be in a positive sense, reducing the effort of a patent examiner, or in a negative sense, increasing the grant of similar patents.
Looking to the future
Of course technology is continually changing, and our past experience in R&D has shown that it is difficult to predict the future. This is why we have a current focus on responsible AI and AI ethics.
Get in touch if you want to discuss this further.
This post was derived from one on LinkedIn.
If you want to stay up-to-date with us on LinkedIn, follow our Page.
