OpenAI has been in the spotlight recently due to its development of a watermarking system for ChatGPT-created text. This system has been in the works for about a year, but the company is facing internal conflict over whether or not to release it to the public.
The Watermarking System
The watermarking system works by adjusting how the ChatGPT model predicts the most likely words and phrases that will follow previous ones, creating a detectable pattern. According to reports, this system has been found to be “99.9% effective” in making AI text detectable when there is a sufficient amount of it. This could be a game-changer for educators looking to combat plagiarism by students who use AI to generate writing assignments.
Public Opinion and Concerns
In a survey commissioned by OpenAI, it was found that people worldwide supported the idea of an AI detection tool by a margin of four to one. However, internal concerns within the company have raised doubts about the potential implementation of the watermarking system. Nearly 30 percent of surveyed ChatGPT users indicated that they would use the software less if watermarking was introduced. Additionally, some employees are worried that the system could be easily circumvented using tactics like translating text back and forth between languages or adding and deleting emojis.
Despite the effectiveness of the watermarking system, OpenAI is at a crossroads regarding its release. While some believe that it is a responsible step to take, others are concerned about the impact it could have on user experience and the company’s financial bottom line. The internal debate reflects the delicate balance between ethics and business interests in the world of AI technology.
In response to the mixed feedback from users and employees, OpenAI is considering alternative approaches that may be less controversial but are unproven. While the company continues to grapple with the decision, it is clear that finding a solution that satisfies both ethical considerations and practical concerns is paramount. Whether OpenAI ultimately decides to release the watermarking system or explore other options, the ongoing debate highlights the complex challenges facing the intersection of technology and ethics.
Leave a Reply