| Evaluation Perspective | Evaluation Dimension | Researcher | Evaluation Approach |
| Source | System Credibility | EU[10] | Accountability, Inclusiveness, Autonomy, Fairness, Privacy, Robustness, Security, Transparency |
| OECD[11] | Sustainable Development, Values, Fairness, Transparency, Explainability, Robustness, Security, Safety | ||
| Singh[12] | Fairness, Explainability, Robustness, Privacy, Security, Appropriateness | ||
| Fujii[1] | Integrity, Robustness, System Quality, Agility | ||
| He Jifeng[14] | Robustness, Self-Reflection, Adaptability, and Fairness | ||
| Channel | Media Credibility | Schweiger [15] | Media Type, Media Subcategory, Media Product, Editorial Unit, Information Creator, Information Presenter |
| Metzger[16] | Source, Information, and Channel Transmitting Information | ||
| Li Xiaojing[17] | Credibility of Media Organizations/Journalists (Source), Credibility of News Reports (Information), etc. | ||
| Zhang Hongzhong and Ren Wujiong[18] | Public Trust in Mass Media, Trust in Social Media, Human-Machine Trust Based on Machine Identity, Human-Machine Trust Based on Language Dialogue | ||
| Content | Content Credibility | Sundar[19] | MAIN Model (From Technological Affordance to Credibility Judgment) |
| Flanagin and Metzger [20] | Dual-Process Model (Heuristic, Systematic) |
| Hilligoss and Rieh [21] | Integrated Framework for Credibility Assessment (Construction Layer, Exploration Layer, Interaction Layer) | ||
| 互层） |
| Objective Layer | Criterion Layer | Sub-criterion Layer | Description of Sub-criterion Layer Indicators |
| AIGC | System Credibility | Training Data Credibility | Dataset Transparency, Source Reliability, Data Robustness, Da |
| Credibility Assessment U | (U1) | (U11) | ta Coverage, Data Scalability, Noise Removal, and Source Security |
| Algorithm Model Credibility $\left(\mathrm{U}_{12}\right)$ | Model Interpretability, Robustness, Stability, Adaptability, Privacy Protection, and Plasticity | ||
| Output Result Credibility (U13) | Accuracy, Error Rate, Precision, Recall, and F-measure of Generated Content | ||
| Media Credibility (U2) | Professional Service Capability $\left(\mathrm{U}_{21}\right)$ | Content Accumulation in Professional Fields, Professional Media Cooperation, Standardized Process Mechanisms | |
| Social Influence $\left(\mathrm{U}_{22}\right)$ | Platform's Social Awareness, Social Evaluation, Third-party Certification and Endorsement | ||
| Neutrality of Stance $\left(\mathrm{U}_{23}\right)$ | Platform's Selflessness, Consistency, Independence, Profitability | ||
| Guarantee Mechanism Construction (U24) | Platform's User Feedback Mechanism, Accountability Mechanism, Privacy Mechanism/Data Security Mechanism | ||
| Content Credibility (U3) | Formal Credibility (U31) | Citation Normativity, Response Consistency, Expression Precision, Source Professionalism, Authoritative Endorsement, Information Noise | |
| Substantive Credibility (U32) | Content Authenticity, Content Correctness, Content Objectivity, Information Timeliness, Information Completeness, Information Consistency, Quality Robustness, Content Security, Content Fairness |
| Relative Importance Level | Definition | Description |
| 1 | Equally Important | The importance of the two indicators is the same |
| 3 | Slightly Important | Based on experience or judgment, one indicator is slightly more important |
| 5 | Quite Important | Based on experience or judgment, one indicator is quite important |
| 7 | Extremely Important | In practice, one indicator is extremely important |
| 6 | Absolutely Important | There is sufficient evidence to show that one indicator is absolutely important |
| 2, 4, 6, 8 | Intermediate Value of Adjacent Judgments | Used when a compromise judgment on indicators is required |
| n | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 |
| RI | 0 | 0 | 0.58 | 0.90 | 1.12 | 1.24 | 1.32 | 1.41 | 1.46 |
| $\mathrm{U{-}U_{k}}$ | $\mathrm{U}_{1}$ | $\mathrm{U}_{2}$ | $\mathrm{U}_{3}$ |
| $\mathrm{U}_{1}$ | 1 | 2 | 1/2$ |
| $\mathrm{U}_{2}$ | 1/2 | 1 | 1/5$ |
| $\mathrm{U}_{3}$ | 2 | 5 | 1 |
| $\mathrm{U_{1}\mathrm{-U_{1i}}}$ | $\mathrm{U_{11}}$ | $\mathrm{U}_{12}$ | $\mathrm{U}_{13}$ |
| $\mathrm{U_{11}}$ | $2$ | $1/2$ | |
| $\mathrm{U_{12}}$ | $1/2$ | $1$ | $1/4$ |
| $\mathrm{U_{13}}$ | $4$ | $1$ |
| $\mathrm{U_{2}\mathrm{-U_{2i}}}$ | $\mathrm{U}_{21}$ | $\mathrm{U}_{22}$ | $\mathrm{U}_{23}$ | $\mathrm{U}_{24}$ |
| $\mathrm{U}_{21}$ | 1 | 1/5 | 1/6 | $1/3$ |
| $\mathrm{U}_{22}$ | 5 | 1 | 1 | 3 |
| $\mathrm{U}_{23}$ | 6 | 1 | 1 | 3 |
| $\mathrm{U}_{24}$ | 3 | 1/3 | 1/3 | 1 |
| $\mathrm{U_{3}\mathrm{-U_{3i}}}$ | $\mathrm{U}_{31}$ | $\mathrm{U}_{32}$ |
| $\mathrm{U}_{31}$ | 1 | $1/7$ |
| $\mathrm{U}_{32}$ | 7 | 1 |
| U | System Credibility $\mathrm{U}_{1}$ | Medium Credibility $\mathrm{U}_{2}$ | Content Credibility $\mathrm{U}_{3}$ | Overall Ranking Weight |
| 0.277 | 0.128 | 0.595 | ||
| Training Data Credibility $\mathrm{U_{11}}$ | 0.286 | 0.079 | ||
| Algorithm Model Credibility $\mathrm{U}_{12}$ | 0.143 | 0.040 | ||
| Output Result Credibility $\mathrm{U_{13}}$ | 0.571 | 0.158 | ||
| Professional Service Capability $\mathrm{U}_{21}$ | 0.065 | 0.008 | ||
| Social Influence $\mathrm{U}_{22}$ | 0.384 | 0.049 | ||
| Neutrality of Stance $\mathrm{U}_{23}$ | (missing value corrected to) 0.401 | 0.051 | Guarantee Mechanism Construction $\mathrm{U}_{24}^{-}$ | |
| 0.150 | 0.019 | Formal Credibility $\mathrm{U}_{31}$ | ||
| 0.125 | 0.074 | Substantive Credibility $\mathrm{U}_{32}$ | ||
| 0.875 | 0.521 | 0.521 |