Developers who have been writing code for decades may soon find themselves in a tricky situation. We know that AI can write code or parts of code and is very good at it if a developer divides the process into chunks. It helps save time and reduces a lot of the tedious work traditionally involved in code writing.
But who owns the AI-generated code?
This is the question that many are struggling to find an appropriate answer to. Where the question is valid, it doesn’t have a simple, straight answer.
But things are more complex than you might think. There are many elements involved that developers can’t sleep over. For example, if the code contains business secrets, the risks are that they have been shared with OpenAI, Microsoft, Google, or any firm whose AI model you used.
This is a likely situation. On an application, you are at work. For the most part, that application is your own effort. You wrote the majority of the code, created the business logic, and specified the user interface. However, you’ve written a few modules using GPT-4, Microsoft Copilot, Google Gemini, or Anthropic Claude and integrated the resulting code into your application.
Who is the owner of the AI-produced code? And does the program as a whole lose any ownership claims you may have made because of the inclusion of that code?
Ownership considerations for complex products
Professionals who specialize in data security and intellectual property point out that copyright and contract law are two different legal disciplines with distinct legal frameworks.
One thing the legal system has been very clear about in the era of artificial intelligence is who owns the content created by AI: since it was not developed by a human, it cannot be protected by copyright.
It isn’t owned by the AI, the people who trained it, or the person who asked it to produce this information. An owner is not possible for that code.
Does any of that code have a copyright? Or is every piece of software we write these days so thoroughly damaged that it may no longer be defended as a piece of intellectual property covered by copyright? In a recent decision, the curatorship of an AI-generated text collection resulted in the author being granted copyright over the collection.
What are the implications for coders in a scenario where the next line might be created by AI and not protected, yet the previous line might have been written by a human and thus protected? It’s a problem any attorney will agree with.
This Pandora box has yet to be sorted out in courtrooms. Commercial firms can definitely not afford to lose the efficiency that they have gained with these AI models, so they might be ready to avoid any red flags raised by the legal teams.
Meeting client requirements at lightning speed will remain a practice until a hefty lawsuit determines the ownership and copyright status of generative AI.
