The urgency to retrofit password authentication systems has grown significantly with the rise of computational power and the integration of advanced Artificial Intelligence (AI) and machine learning models in password guessing tools. As these tools become more sophisticated, creating passwords that are both strong enough to withstand such challenges and memorable has become increasingly difficult. This difficulty often leads users to adopt unsafe practices, such as reusing passwords or creating weak passwords. The RoseCliff Algorithm offers a promising solution by introducing a dual authentication mechanism that not only fortifies systems against sophisticated guessing techniques but also dynamically updates stored passwords. Our research demonstrates that this algorithm significantly increases the time required for a hacker to decipher a user’s password using current dominant hacking techniques. Additionally, we show that large language models (LLMs) can assist in generating stronger memorable passwords and creating more effective honey tokens. We propose that leveraging LLMs, such as ChatGPT, in conjunction with the RoseCliff Algorithm, could pave the way for the next generation of robust and dynamic authentication systems, offering enhanced security and usability for users.