Analysis Effect of Video, Voice and Image Abuse Through AI Generator in Indonesia

Many people agree that technology can help us in creating content, one of which is AIGC. AIGC helps users in creating content more efficiently and attractively. If we put aside the positive side, AIGC can also cause negative things, such as manipulation of images, videos, or audio to commit crimes. This study aims to analyze the risks posed by AIGC, such as deepfake with the aim of committing crimes. There are 7 variables that will be analyzed in this study, including Perceived Usefulness, Emotional Risk, Information Quality, User Trust, Social Influence, Behavior Intention, and User Behavior which create 6 hypotheses. This study analyzes how perceived usefulness, emotional risk and information quality factors can affect user trust, then how social influence factors affect behavioral intention, and how user trust factors can affect behavioral intention and user behavior. Researchers use a sampling method which means that respondents are only intended for demographic conditions that are wary of AIGC content and are in Indonesia. The data was collected by distributing questionnaires using e-forms and taken from May 8, 2025, to May 11, 2025. The results of this study were taken based on 475 respondents with the majority aged 12-27 years and dominated by female gender, and the results showed that the Perceived usefulness and information quality factors have a significant influence on user trust. In addition, the Social Influence factor also has a significant influence on user behavioral intention, and the user trust factor also has a significant influence on user behavioral intention and user trust to access AIGC content. Meanwhile, only the emotional risk factor does not have a significant influence on user trust.多数人认为，技术能够助力内容创作，其中代表性方向之一便是人工智能生成内容（Artificial Intelligence Generated Content，简称AIGC）。该技术可帮助用户更高效、更具吸引力地完成内容创作工作。暂且搁置其积极面向不谈，人工智能生成内容也可能引发各类负面影响，例如通过篡改图像、视频或音频内容实施犯罪活动。本研究旨在分析人工智能生成内容所带来的各类风险，例如用于实施犯罪的深度伪造（deepfake）技术。本研究将分析7项核心研究变量，涵盖感知有用性（Perceived Usefulness）、情感风险（Emotional Risk）、信息质量（Information Quality）、用户信任（User Trust）、社会影响（Social Influence）、行为意向（Behavior Intention）与用户行为（User Behavior），并据此提出6项研究假设。本研究将探究感知有用性、情感风险与信息质量等因素对用户信任的影响路径，同时分析社会影响因素对行为意向的作用机制，以及用户信任因素对行为意向与用户行为的影响关系。 本研究采用限定性抽样方法，调研对象仅涵盖印度尼西亚境内对人工智能生成内容持警惕态度的人群。数据通过电子问卷发放形式收集，采集时段为2025年5月8日至2025年5月11日。本研究共回收475名受访者的有效数据，其中受访者年龄多集中于12至27岁区间，且女性受访者占比更高。研究结果显示，感知有用性与信息质量因素对用户信任具有显著正向影响。此外，社会影响因素对用户行为意向具有显著正向影响，用户信任因素对行为意向以及用户访问人工智能生成内容的信任水平均存在显著影响。而唯有情感风险因素对用户信任未产生显著影响。