Impact of Artificial Intelligence-Generated Content Labels On Perceived Accuracy, Message Credibility, and Sharing...

Area of Science:

• Artificial Intelligence
• Human-Computer Interaction
• Digital Media Studies

Background:

• Generative AI and AI-generated content (AIGC) are increasingly prevalent on virtual platforms.
• Social media platforms are beginning to label AIGC, but research on their effects is limited.

Purpose of the Study:

• Investigate the impact of AIGC labels on perceived accuracy, message credibility, and sharing intention for misinformation.
• Refine the strategic application of AIGC labels in digital environments.

Main Methods:

• A 2x2x2 mixed experimental design was employed with AIGC labels (present/absent), information type (accurate/inaccurate), and content category (for-profit/not-for-profit) as factors.
• 400 participants in both experimental and control groups evaluated content, providing feedback on accuracy, credibility, and sharing intention.
• Statistical analyses included repeated-measures ANOVA and simple effects analysis.

Main Results:

• AIGC labels did not significantly affect perceived accuracy, message credibility, or sharing intention.
• Information type and content category significantly impacted all three dependent variables.
• Significant interaction effects were found between information type and content category for accuracy and credibility, and between information type and AIGC labels for sharing intention.

Conclusions:

• AIGC labels minimally influence user perceptions of accuracy and credibility but aid in distinguishing AIGC from human content.
• Labels do not negatively affect user perceptions of platform content, suggesting potential for fact-checking and governance.
• Nuanced AIGC labeling strategies are necessary, varying by information type, particularly for misinformation where labels may slightly enhance sharing intention and perceived accuracy.