Seeing is Reading: VLM Intensive with Carry AI

Today’s AI workshop session focused on Carry AI’s development of VLM (Vision language model), which can effectively identify human actions, read human expressions and recognise objects to anticipate potential threats. The founder of CarryAI, Samson Li, hosted the session and presented VLM’s immense capacity to be assimilated upon society. From CCTVs and IP camera streams, to automatic scanning of illegal activities, including more practical uses such as mechanical checkups of construction materials for workplace safety, our host demonstrated fully the integration of VLM into everyday life.

Today’s experience expanded our understanding of AI's ability to address real-world challenges that traditionally demand human labour and time. Moreover, it inspired us to reconsider the countless applications of AI, moving beyond merely generating ideas to the practical functionalities of technology. Additionally, the live demonstration of Carry AI’s usage, such as how it identified team members playing cards as “gambling”, showed all of us the distinct prompts and configurations required for different functions.

While technology has accelerated at a remarkable pace, Mr. Li also emphasized the importance of human engagement, precision and intentionality to reap the benefits of AI. This enhanced our technological literacy as we navigate the mechanisms needed for Carry AI to perform complex tasks, deepening our knowledge of what lies behind the complicated systems that we often overlook. Ultimately, we were deeply fascinated by Mr Li’s presentation and felt encouraged to approach AI as an invaluable resource that deserves to be explored beyond theoretical discussion. We also were enlightened to not become overeliant on it as individual cogitation can never be replaced.