Using AI to generate new IP

Artificial Intelligence (AI) is what we consider a ‘swiss army knife’ technology. It has so many different applications across different business functions, that ‘boxing’ it to a few use cases overlooks the range of capabilities AI can provide.

One of these areas is in the creation of new intellectual property (IP).

AI has already demonstrated a capability for invention. Given a clear set of parameters and a goal, AI can test millions of potential configurations and, with human support, arrive at novel solutions. This has wide applications, whether it be designing a toothbrush or bicycle, testing medical treatments, or calculating solutions to difficult math problems.

This provides massive opportunity for organisations that profit from IP. It provides a relatively low cost way to explore options and narrow the viable candidates before humans get involved.

As such it’s worth noting AI can’t own IP for its discoveries. This has been tested in a number of countries, as documented in The Conversation, Artificial ‘inventors’ are pushing patent law to its limits.

This article particularly looks at a test case where a company attempted to register a patent for a new food container that was solely created by an AI named DABUS (Device for the Autonomous Bootstrapping of Unified Sentience). This has been rejected by countries including Australia, the United Kingdom, United States, the European Union, Germany, South Korea, Taiwan and New Zealand.

In South Africa the patent was provisionally granted, but has not yet been substantively examined.

In practice this means a human must be named as the inventor in any patent, even if the work was done by an AI.

It also means if your organization uses an AI supplied by a third party, their ability to claim ownership over IP created by the AI is limited. Your organization is the beneficiary and can name the inventors, humans who defined the problem or contributed to the solution, such as by verifying the AI’s solution.

We see the ability for AI to take existing works – including corporate knowledge – and produce new IP, every day in our business.

While it is not guaranteed AI will develop valuable IP for an organization, if you have a solid base of corporate knowledge there is a good likelihood AI can link concepts to form unique new value.

So it’s time to think about what valuable IP is hidden in your knowledge waiting for AI to discover.