Thanks for following us on GPTZero’s development! This week, we’re trying out something different — a case study from our team, on an AI model and an exploit, that will hopefully provide unique insights on the AI detection space, and whether it is in fact, as some claim, an ‘artificial intelligence arms race.’
I heard that a popular workaround for AI detectors is taking an AI-generated text and running it through a spinbot (paraphrase software), and then it will pass as human. Is that true? What is GPTzero doing about that, or can anything be done about that?
>> We’re also excited to be entering into collaborations this week with some of the largest Learning Management Systems, to build the best AI detection solution for teachers.
Please tell me it's Canvas. It would tremendously helpful if I could open a student's exam (not just essays, but also short/open-ended answers) and click a 'check GPTzero' button. Or better yet, have it automatically check everything for me and display an alert when it senses AI-ness.
It is clear that AI detection requires continuous work to ensure that the models are up-to-date and effective. It is good to see that GPTZero is taking a proactive approach to monitoring and responding to new AI exploits. It would be interesting to see if the "AI arms race" is a real phenomenon or a result of current AI models. Either way, it is worth noting that AI detection requires an iterative approach that constantly monitors and reacts to new exploits.
One question: is GPTZero built for English only? Everything I write and test in Swedish passes as human written when I it's really not