A preliminary validation scale for general attitudes toward artificial intelligence

Exploring AI's impact through expert-driven studies and comprehensive surveys.

Our Mission

We conduct innovative research to assess AI's impact through rigorous mixed-methods studies and expert collaboration.

AI-Driven Insights

Expertly crafted surveys for comprehensive analysis of cognitive, affective, and behavioral dimensions in AI.

Item Generation Process

Utilizing AI to create and refine survey items for accurate psychometric evaluation and cultural adaptability.

Pilot Survey Analysis

Engaging diverse respondents to ensure robust data collection and insightful analysis across distinct regions.

Research Study

Exploring AI's impact through mixed-methods and expert analysis.

Phase One

Item generation and expert review for content validity.

A poster features a close-up of a woman's face with a message about psychology. The top of the poster has the phrase 'GET PSYCHED!' in large orange text. Below, there is a series of questions in varying sizes and colors, exploring different psychological topics. The bottom part of the poster contains a web address.

Phase Two

Pilot survey and analysis to refine generated items.

With the rapid advancement of AI technology and its widespread adoption in business, healthcare, education, and public services, public attitudes toward AI have become increasingly multifaceted. On one hand, people hold high expectations for AI-driven efficiency gains, intelligent services, and innovative solutions; on the other hand, concerns about privacy breaches, job displacement, algorithmic bias, and threats to human autonomy persist. Existing attitude scales tend to focus on specific applications or technologies, lacking a comprehensive instrument to assess overall perceptions, emotions, and behavioral intentions toward AI. Therefore, this study aims to develop and preliminarily validate a “General AI Attitude Scale” (GAIAS), which will include items across three dimensions—cognition (Perceived Usefulness, Perceived Risk), affect (Trust, Anxiety), and behavioral intent (Usage Intention, Resistance Intention)—and employ Exploratory and Confirmatory Factor Analysis along with Item Response Theory (IRT) to examine the scale’s factor structure and item properties. Our central research question is: Can we leverage the GPT-4 API to generate and refine culturally sensitive, psychometrically sound scale items, and validate their factor structure, reliability, and validity through empirical surveys?