User-centric AIGC products: Explainable Artificial Intelligence and AIGC products

Users struggle with AIGC tools because they don't understand how these systems work. The paper identifies "user experience challenges" with generative AI tools like ChatGPT and Midjourney, suggesting that a key barrier to effective use is the "black box" nature of how these systems generate outputs.

Explainable AI (XAI) can make generative tools more user-friendly. By integrating interpretable AI methods into AIGC products, users can better understand why the system produced certain results, which the authors argue will "enhance the user experience for ordinary users."

The solution requires focusing on two critical moments: input and feedback. The paper proposes that XAI should be integrated specifically at the "input and adjustment feedback stages" of AIGC products, meaning users need clarity both when they're giving instructions and when they're refining results.

Human-centered design is essential for responsible AI development. Rather than building AI tools and expecting users to adapt, the paper advocates for a "human-centric design approach" that prioritizes user understanding and control—a foundational principle for responsible AI deployment.

Current AIGC products lack transparency that ordinary users need. The emphasis on making XAI accessible to "ordinary users" (not just AI experts) highlights that most people using these tools today cannot understand or predict how they work, creating a gap between capability and user comprehension.