Using AI to Write Your Code May Mean You Don’t Own It

One of the great uses of AI is writing code. It’s fast, it’s easy, and it’s elegant. It’s also likely not yours. 

Ownership is a legal construct. You cannot own anything without a law that sets out ownership rights. That is as true for physical property as for intellectual property (“IP”). IP includes patents, trademarks and copyrights. In the US, computer code is covered under copyright law, US Code Title 17, and it says that copyright protection is granted if the work is original and the author is human. 

The Copyright Office issued guidance last year around the use of generative artificial intelligence (generative AI) and its impact on copyright. The guidance specifically stated that the use of prompts to generate a work, for example, an image, or computer code, could mean that the generative AI tool would be the author and thus would not receive protection, if the user does not “exercise ultimate creative control over how such systems interpret promotes and generate materials.” The Office likened the use of AI in this manner as providing instructions to a commissioned artist. 

Companies are implementing AI policies, including the use of AI to write code, for various reasons. One is that products and source code are typically highly confidential and should not be provided to third parties. The other is that if your company does not own its code, it may have an impact on the company valuation, particularly if it’s a tech startup. Investors and potential buyers always ask if you own your technology. Answering ‘no’ to an important part of your tech stack will be an issue.

If you are using AI to fix bugs in your code, however, the answer may be different. The analysis is done on a case-by-case basis and what, how, and how much, are factors that inform the answer. 

When looking at whether or not to use AI to assist you with writing – or fixing – code, it’s important to consider several factors, such as how much is being written by AI, the importance of the product you are working on in the tech stack, and the confidential nature of what you are providing to the AI tool. And, of course, always read your company’s AI policy.