Writing Code With AI Poses a Legal Risk

AI is changing software development, but what legal risks come with it? Intellectual property and open source are in focus as AI-generated code comes under scrutiny. How should companies protect themselves against infringement in an era where AI writes the code?

Share
Writing Code With AI Poses a Legal Risk

AI-powered coding agents are poised to revolutionize software development, but this innovation comes with legal challenges, particularly around intellectual property. Unlike today's AI assistants, which help programmers write snippets of code, the more advanced AI agents are expected to handle entire blocks of code and could potentially replace roles like junior developers.

Tech-focused lawyers warn that organizations need to carefully review the code AI agents produce to avoid potential infringement of copyrighted work.

Intellectual Property Risks in Focus

Media outlets and creators have already sued AI vendors for using their material in AI models without permission. Similar problems could arise in the software industry, even though enterprise source code is rarely shared openly. Jeffrey Gluck, an attorney at Panitch Schwarze, emphasizes the risk that AI generates code reflecting copyrighted data from its training material.

How the AI is trained becomes crucial. If the training data includes copyrighted code, the generated code risks violating laws and license agreements.

Open Source Creates Further Dilemmas

Many AI models are trained on open-source code, which is often restricted to non-commercial use. This creates a gray area when AI-generated code is used commercially, since license terms risk being violated, according to Ilia Badeev of Trevolution Group.

AI vendors like GitHub have taken steps to reduce the risk. GitHub Copilot, for example, offers a code reference filter to identify and block suggestions that match public code. They also have an indemnification policy for users who enable this protection.

But despite these efforts, GitHub and other players like Tabnine place the responsibility on users to ensure their license terms are followed.

Code Review Remains Important

To minimize legal risk, organizations need to review AI-generated code the same way they review human-written code. Ilia Badeev points out that best practices for code review remain crucial, even with AI.

New services for reviewing AI-generated code could become a solution to the problem. Platforms that scan code for copyright and license infringement are expected to grow as AI coding agents see wider use.

The Future of AI Coding Remains Uncertain

It's unclear how significant the consequences of legal disputes could be for companies using AI coding agents. While major AI vendors will likely be the primary targets of lawsuits, user organizations aren't immune, especially if their AI-generated code is used in commercially successful products.

Organizations are urged to act proactively by putting safeguards in place to reduce the risk of infringement. The question is: will companies adapt, or turn a blind eye to the risk?