Purpose Endoscopists have increasingly utilised Artificial Intelligence (AI) in recent years to enhance adenoma detection rates. Endoscopy presents a steep learning curve, requiring novice endoscopists to acquire numerous cognitive and motor skills during hurried procedures. Studies have examined the perceptions of experienced endoscopists regarding the use of AI in endoscopy, but have not examined those of novices. This study aims to survey novice endoscopists' attitudes towards knowledge and utilisation of AI in their training. Methodology An online survey was conducted among novice endoscopists from three tertiary institutions in Singapore who have done fewer than 200 procedures. The 27-item questionnaire, adapted from validated surveys, covered four thematic domains. Responses were analysed using quantitative and thematic approaches. Results The majority of the 50 endoscopists surveyed found it challenging to acquire endoscopy-related cognitive and motor skills. Participants believed they were familiar with AI and Large Language Models (LLMs). However, only a small percentage genuinely understood how different algorithms functioned. Many (64.0%) reported limited exposure to AI but expressed interest in understanding it (86.0%). Some agree that AI has improved endoscopy quality (48.0%) and will aid in developing skills (56.0%). Concerns remain regarding accuracy (16.0%) and potential deskilling (20.0%). Conclusion In conclusion, novice endoscopists have limited exposure to AI during colonoscopy training but are open to learning and utilising it further. They remain optimistic about AI’s role as an adjunct to enhance endoscopic training and patient outcomes. However, reservations about deskilling and program accuracy present barriers to fully integrating AI into colonoscopy training.