Friends, Secrets, and AI: The Tension Between Digital Trade Secrets and Transparent AI in the Age of ChatGPT

Abstract

<p>Friends are an integral part of our lives, and one of the hallmarks of a good friend is their ability to keep our deepest secrets. By comparison, digital trade secrets are a good friend for ChatGPT because they keep its inner secrets (algorithms and source codes) legally under lock and key. The friendship is increasingly coming under scrutiny as the public is now demanding a reasonable explanation of how this technology works to build trust in its outputs. This has ignited a debate regarding the balance between transparent AI and the protection of digital trade secrets. The legal basis for digital trade secret protection derives from international agreements, such as the Agreement on Trade-Related Aspects of Intellectual Property Rights (TRIPS), as well as regional frameworks like the EU’s Trade Secret Directive (TSD). These frameworks provide for the protection of digital trade secrets while acknowledging the need for transparency in certain situations. Yet, attaining meaningful transparency in ChatGPT which operates on a closed source policy, is a daunting task. However, the demands for transparency are only expected to grow with the coming on the market of DeepSeek which purportedly operates based on open-source policy. At this point, two questions need to be addressed. What is meant by transparent AI, and is there such a “thing” as a “digital” trade secret?<br></p>