For almost a century, social scientists have relied on two main methods for gauging attitudes: Thurstone scales and Likert scales. Thurstone scales, known for their precision, involved a complex process of gathering expert judgments, which made them less popular than the simpler Likert scales. But what if technology could eliminate the difficulty of creating Thurstone scales? And what if we could use AI to make the process even more objective? A new study explores exactly this by using a Large Language Model (LLM) to help develop a Thurstone scale measuring attitudes towards people with AIDS. Researchers compared the LLM's categorization of attitude statements with those of a diverse group of 75 human judges, from students to seasoned social researchers. The AI didn't just categorize; it explained its reasoning, offering a fascinating glimpse into how machines process language and sentiment. The results? Remarkable agreement between the AI and human judges on a majority of the statements. This suggests that LLMs could play a valuable role in creating more reliable and less biased attitude scales. This research opens exciting doors for using AI to improve how we measure attitudes, potentially leading to more accurate and insightful social research across various fields. Future research will focus on developing specialized algorithms within LLMs, fine-tuned for the nuances of attitude measurement, promising even greater accuracy and efficiency in understanding human attitudes.