OpenAI has created a new tool that is “very accurate” in recognizing AI-generated content that was produced by ChatGPT. Be that as it may, the company opted out of releasing it for two reasons, i.e. fears bogus and stigmatization of non-English speakers.
Internal Debates Stall Tool Release
OpenAI reported in a June update on a blog post that had been published in May of the previous year it was inter-company debates this time that had delayed the introduction to the market of those attractive new detection devices. A Wall Street Journal report revealed these issues on the 4th of August, presuming that the company is concerned about the broad-reaching effects of such technology.
Concerns Over Bypassing Detection
Despite the accuracy of the tool and its resistance against localized alterations OpenAI concedes that the system is not immune to adversities. The company is apprehensive that cyberbullies might create a technology “backdoor”, so to speak, that would defeat the detection mechanism and thus consequently the product would be misused.
Potential Stigmatization of Non-English Speakers
The company particularly worries about the way such software is going to affect non-English speakers. The findings of a study indicate that the watermarking method may worse hit these language users. The corporation is worried that it might scare away non-native English speakers with the utilization of AI as the writing instrument, because of an exploit that translates English text into other languages to avoid detection.
“Another important risk we are evaluating is our research showing text watermarking can be less fair to different groups. It could, for example, lead to stigmatized usage of AI as a helpful writing tool for non-native English speakers” OpenAI said.
Lack of High-Accuracy AI Detection Tools
At present, AI content detection tools are not the case on the market that provide the greatest accuracy across general tasks, especially in refereed research. OpenAI’s tool, which depends on unnoticeable watermarking and exclusive detection techniques, would be one of its resources, developed in-house for the company’s models differently.
The Future of AI Content Detection
While the approach of the OpenAI company not to unveal the instrument that can recognize fake news seems to be rather conservative, it shows the complexity and etical issues of the subject. The company is still going on about using AI to optimize the processes that positively or negatively interfere with societal improvements.
With the debate surrounding AI-generated content and its detection moving to new territories, it is not yet certain whether such tools will be released or not. Qualities of OpenAI remain to be very careful when it comes to putting forward new technologies, criteria such as fairness and trust also make up the company’s goals.