False Alarms: AI Errors Cost Writers Their Jobs 

With the advent of ChatGPT, organizations started to search for ways to catch people who were trying to pass off AI generated text as their own writing. Several startups were launched to detect the use of AI including, Copyleaks, GPTZero, OriginalityAI, etc.  

These companies were launched to account for the authenticity of written data. Some advertised accuracy rates to be as high as 99.98%. The present situation, studies and experts suggest otherwise, considering the number of false results given by these detectors which in turn created trouble for a lot of people.  

Kimberly Gasuras, a local journalist in Bucyrus, Ohio, who has been a news reporter for 24 years was accused of using AI and was rid of her job.  

She used to work for a freelance writing company, WritersAccess, where she wrote blogs and other content for small and medium enterprises. Her income plummeted as some clients switched to ChatGPT for their writing needs, making it a difficult time for her. Soon she was warned alleging that she had used an AI tool called ‘Originality’. Despite her attempts to deny the claims and even after being cautious, using AI detectors herself before submitting to avoid any confusion, she was still fired saying that her account showed excessive use of AI.  

“This technology doesn’t work the way people are advertising it”, said Bars Juhasz, co-founder of Undetectable AI, which helps people to humanize AI text to sneak it past detection software. He adds that there are several concerns regarding the reliability of these AI detectors even though they claim to be 99% accurate. He says that even if it’s true, there still could be a false positive in every 100 tests, which could affect the livelihoods of people and their reputation.  

Even though the academic world was the first to adopt the use of AI detectors to identify plagiarism, the false accusations pushed many universities to ban the use of AI detection software.  

With Google tracking down AI generated content to clean up the online junk, it is understandable for the organizations to be using AI detectors to identify computer generated data, but the false alarms that cost people their livelihood and reputation is worrisome.