Generative AI is capable of many things which we are becoming increasing aware of day by day. To manage the use of Generative AI, some institutions are implementing traffic light systems which categorises generative AI usage: Red = no AI use, Amber = limited AI use under guidance, Green = AI use permitted within guidelines.

Despite several institutes adopting this type of strategy which is similar to or based on the AIAS by Perkins et. al (2024), there a couple of issues with this type of strategy:

For the red-light category, as an academic, how do you detect AI use reliably? Currently, AI detection tools are not reliable and a simple defence is all that is required that "I asked Generative AI to help me rewrite my paragraphs". For any middle categories such as amber or similar, again, how can you detect its use?

For generative AI usage, the traffic light system doesn't provide a solution to assessment in s generative AI era in higher education. A strategy which is more logical in this era is the two-lane approach to assessment (Bridgeman and Liu, 2024). This categorises assessment types based on the assessment security: Secure (in person) or Open (unsupervised).