The widespread use of ChatGPT for cheating on online exams has become a major concern. Many exam platforms only prohibit text copying from their webpages, which can be easily bypassed through Chrome's web console. Developing a more secure platform that prevents copying questions (to provide it into chatGPT) through various methods is essential. However, with the upcoming advancement of ChatGPT's image understanding capabilities, individuals may soon be able to simply take a photo of a question and receive the answer. So any technical protection of the webpage become useless. One potential solution involves having exam-takers enable their webcams for live monitoring by a proctor. This method, while effective for some high-stakes exams, is not cost-effective or scalable for mass-market use. The challenge lies in finding a way to protect exams from ChatGPT and the "photo-of-screen" approach without relying on human intervention. What do you think? Is there a way? Or a human control is the only option? -- PS: Text is optimised and shortened with ChatGPT of course ))