Title page
Contents
Overview and key findings 5
1. Introduction 7
2. Approaches to measuring false and misleading content online 8
3. The Truth Quest methodology 12
4. Truth Quest findings 21
5. Conclusion 36
Annex A. The OECD taxonomy of false and misleading content online and its implementation in Truth Quest 37
Annex B. Statistical tables 39
References 42
Endnotes 46
Table 1. Truth Quest country coverage and languages 13
Table 2. Truth Quest score by confidence level 24
Table 3. Perceptions of AI and overall Truth Quest score 31
Table 4. Truth Quest score and trust in news from social media 35
Figure 1. OECD taxonomy of false and misleading content online 14
Figure 2. Truth Quest interface: Instructions and avatars 19
Figure 3. An example of the Truth Quest frame 20
Figure 4. Ability of adults to identify the veracity of online news 22
Figure 5. People's perception of their ability to recognise false and misleading content online 23
Figure 6. Average Truth Quest scores by type 24
Figure 7. Truth Quest scores by theme 26
Figure 8. Truth Quest scores for AI- and human-generated true claims 27
Figure 9. Truth Quest scores for AI- and human-generated disinformation 28
Figure 10. AI- and human-generated content by theme 29
Figure 11. Perceptions of AI and Truth Quest score for AI-labelled content 30
Figure 12. Media consumption patterns 32
Figure 13. Consumption of news on social media 32
Figure 14. Truth Quest score and percentage of adults who often get news from social media 33
Figure 15. Trust in news sources 34
Figure 16. Trust in news from social media 35
Boxes
Box 1. The psychology associated with labelling content online 16
Annex Tables
Table A B.1. Population and quotas used for targeting 39
Table A B.2. Matrix of claims in the Truth Quest database 40
Table A B.3. Key behavioural and perception-related questions in Truth Quest 41
Annex Figures
Figure A A.1. Decision tree to categorise the Truth Quest claims 38