The rapid development and widespread use of Artificial Intelligence (AI) tools has brought a range of benefits to businesses and individuals. However, increased access to AI tools also means unprecedented levels of misinformation, with deep fakes presenting a particular concern globally. Recognized as the number one global threat by the 2024 World Economic Forum’s Global Risks Report, AI misinformation poses a serious problem for the world.
While the issue is global, certain countries seem to be grappling with deep fake content more so than others. A study by DailyAI explored this phenomenon, analyzing the frequency and depth of interest in deep fake-related tools in countries leading in AI research. The results revealed that Singapore, Australia, Finland, and the Netherlands have the highest interest in the creation of deep fakes. Interestingly, China, despite being the second-ranking nation in AI, is the least interested in AI overall according to the findings.
Australia was found to have the highest interest in deep fake creators, with an occurrence rate of 0.41 out of 100,000 people. The country also had a significant number of searches for “deep fake porn maker” and “deep fake image maker”. In late 2023, the Australian government expressed plans to pressure social media platforms into cracking down on deep fake images and videos, citing growing concerns over online offensive content and hate speech.
Singapore, on the other hand, had the highest frequency of searches for “deep fake porn makers”, despite middling interest in general “deep fake makers”. There is also significant interest in “deep fake image maker”, second only to Finland. Notably, a deep fake involving Deputy Prime Minister Lawrence Wong was used for an investment scam in Singapore in 2023.
The Netherlands leads the way in terms of searches for deep fake video creators, with almost double the rate of second-ranking Singapore. In contrast, Finland has the highest interest in deep fake images and nudes.
The US, a leader in AI research, surprisingly ranks low in interest toward deep fake creation. However, it is actively drafting laws to mitigate the harm caused by deep fake content. Lastly, despite its prominence in AI, China shows the least interest in deep fake production across all search metrics, possibly due to limitations in accessing accurate search data from China.
While the accessibility of deep fake technology proves concerning, it is possible to discern real content from fakes by focusing on details such as lighting and shadow inconsistencies, facial feature anomalies, mismatched audiovisuals, unusual skin texture, background inconsistencies, and the quality and resolution of the content. By utilizing these indicators, one can identify and caution against deceptive deep fakes.