Will AI generated student essays become undetectable by current plagiarism tools at over 50% false negative rate by 2027?

Advances in AI language models have been rapid and significant, with tools like OpenAI's GPT-4 demonstrating capabilities that may make detected human-like responses more indistinguishable from student essays. Given current trajectories in AI development, including ongoing improvements in generative models and their ability to incorporate stylistic nuances, it is plausible that by 2027, a substantial percentage of these AI-generated texts will evade plagiarism detection, especially considering that existing tools are already facing challenges in accurately recognizing AI-written content.