Generative AI – What Are The Implications For Intellectual Property?

Overview

Generative AI tools are transforming the landscape of how businesses operate. While traditional AI has been around for some time, generative AI goes a step further and creates something new based on the question posed and the datasets it has been trained on.

Below we discuss the questions that businesses need to ask themselves when considering the implications generative AI:

What information does the AI tool use to generate outputs?

Generative AI tools (“AI tool”) can use a range of different sources to train its AI tool including scraping the internet for publicly available information and data inputted by users. At present, there is little regulation on text and data scraping so AI developers are able to use a lot of information/data at will to train AI to be a relatively sophisticated tool.

Getty, an image licensing service, is suing AI platform Stability AI, for doing just that, alleging improper use of its photos by using its image catalogue to build its AI model, which will actually compete with Getty. Businesses should check the particular AI tools terms of use to verify how the AI tool has been trained and what information could be included in the outputs.

What does the AI tool do with the data I input and do we have authority to input the data into the AI tool?

If using third party or client data, you may not be clear on the extent of the permissions you have to use their data through an AI tool, particularly where such data is reused by the AI developer to train the AI tool and re-generate the data in future outputs. The risk is that the third party data is used by others, without an appreciation of its source or ownership. This is an important issue, addressed by OpenAI who recently changed their terms of use for ChatGPT to inform users that any data inputted will not be used to train its tool. This is a change enacted to put users at ease that their information will not be widely disseminated. Businesses looking to incorporate AI tools into their internal processes should ensure they inform clients of how their data may be used.

Who owns the AI outputs of the data?

This is a question of ongoing concern for regulators in most jurisdictions. In the UK, “computer-generated works” are owned by “the person by whom the arrangements are necessary for the creation of the work undertaken.” For the moment, AI generated works (subject always to the terms of use of the particular AI tool used) are assumed to fall within this definition, but this could change as generative AI tools develop and the issues around ownership and subsistence of AI-generated works are litigated. Businesses receiving outputs from AI tools will generally own the IP in those outputs, and therefore may be responsible if those outputs infringe third party intellectual property rights.

What are the pros and the cons of generative AI?

The advantages for businesses using generative AI include increased productivity and the ability to provide personalised and customised outputs, but this must be weighed against the risks generative AI poses, including lack of transparency over how outputs are generated, privacy and security risks of data along with the risk of being pursued for infringing on someone else’s intellectual property rights.