Context: In software engineering (SE), the systematic mapping study (SMS) is one of the methods adopted for evidence-based decision-making, selecting and synthesizing relevant literature on a specific research topic. Tool support is beneficial due to the time-intensive nature of the SMS process and its activities. Gap: Large language models (LLMs) such as ChatGPT-4.o can potentially accelerate repetitive activities, such as the data extraction in the SMS process. Therefore, having a tool to assist this activity could save time and effort. This proof-of-concept study evaluates how ChatGPT-4.o can support SMS activities in SE, particularly data extraction. Method: We assessed the accuracy of utilizing ChatGPT-4.o for extracting data in one SMS, in contrast to the manual extraction. Results: The accuracy of ChatGPT-4.o was 87.83%. Conclusions: Our preliminary findings suggest that entirely replacing the human extraction process with ChatGPT-4.o is not recommended. However, it is promise employing ChatGPT for semi-automated data extraction for evidence syntheses in SMSs in SE.