Abstract
In the context of generative AI’s rapid advancement across various
industries, training models require large and comprehensive datasets to
improve AI models, yielding more accurate and realistic outputs. This
increased demand has reshaped the landscape of data accessibility and
economics, particularly with Application Programming Interfaces (APIs).
This shift has led to data providers and social media platforms
enforcing new access restrictions. Such changes have created significant
barriers for researchers, especially those in social science research,
in acquiring data. This article addresses these issues by evaluating
alternative data collection methods, focusing on their application in
social research. It critically examines the strengths and weaknesses of
these methods, underscoring their practicality and reliability. As AI
continues to transform industries, this paper provides a vital guide for
researchers, data analysts, and businesses to navigate the evolving
dynamics of data collection, particularly in the context of social
research.