While the proliferation of AI tools and technology has already delivered numerous benefits to individuals and businesses alike, it’s also true that the increased accessibility to AI tools has had harmful consequences on the level of misinformation put out into the world.
The use of deep fakes to misrepresent influential figures is a growing concern worldwide – with misinformation and disinformation ranked as the number one threat faced by the world in the next two years by the World Economic Forum’s Global Risks Report (2024).
AI misinformation is a global issue – but which countries are battling these issues most fiercely? In this study, our team at DailyAI took the leading artificial intelligence countries (determined by research capacity) and analyzed the average frequency and depth of interest in deep fake-related tools.
Key Findings:
Sam Jeans, AI expert of DailyAI, comments:
“The steep rise in deep fakes and manipulated content posted online has understandably been a huge cause for concern for governments around the world, as globally we try to find the correct boundaries between harnessing new and emerging technologies and fighting against the things that can cause unparalleled harm.
These findings demonstrate that interest in deep fake manipulation is certainly more prevalent in some countries than others, but with this content accessible at a global scale, a group effort will be needed to ensure the harm caused by deep fake content is minimized.”
When assessing the top 10 artificial intelligence countries for their interest in deep fake makers and related searches, we found that Australia stands out as the most interested, with an occurrence rate of 0.41 out of 100,000 people.
Australia also features within the top three countries with the most searches for “deep fake porn maker”, and fourth for “deep fake image maker”. Australia is currently considered to be the sixth top leading country for artificial intelligence capabilities and research.
At the end of 2023, Australia announced that they would be pressuring social media platforms to crackdown on deep fake images and videos after growing concern over the level of offensive content and hate commentary being put out online.
Our analytics also showed that Singapore ranked highest overall for searches for deep fake porn makers, with an occurrence rate of 0.50 per 100,000 people. In addition, the country is second most interested in deep fake video makers, and joint-second most interested in “deep fake nude makers”. Searches for this term have an occurrence rate of 0.17 per 100,000 people – the same as the Netherlands and the US.
This finding is somewhat surprising considering Singapore’s interest in general “deep fake makers” is far lower – with the country sitting in the middle of the ranking for that search term.
Singapore is, however, also the second-most fascinated by the search term “deep fake image maker”, with an occurrence rate of 0.17 per 100,000 people. The country ranks second only to Finland for this search term.
Singapore is ranked as the third leading country for artificial intelligence research and capabilities. One of the most notable occurrences of deep fakes used in Singapore was an investment scam video which circulated on social media in 2023, using the Deputy Prime Minister, Lawrence Wong, and a fake voice-over promoting the scam investment opportunity.
While the Netherlands does not feature particularly highly for other deep fake-related searches, the country has an undeniable interest in making deep fake video content in particular.
The occurrence for searches for deep fake video makers was 0.96 per 100,000 people – almost double that of the country ranking in second – Singapore, again, at 0.50 searches per 100,000 people.
The Netherlands is another country in which deep fakes of politicians have been used to mislead public opinion. In the November 2023 elections, an alarming number of deep fake manipulations were reported to be circulating on social media of numerous prominent politicians.
Finally, Finland ranks amongst the top countries most interested in deep fakes, with the highest occurrence of searches for deep fake nude makers and image makers (0.18 per 100,000 people for each respective search term).
Finland is currently ranked as the ninth leading country for artificial intelligence research and capability. The country is often applauded for its mission to put education at the forefront of the battle against misinformation and fake news – but as demonstrated here, that doesn’t always deter people from being interested in what these new forms of AI can achieve.
The US is currently considered to be the leading country in artificial intelligence research and capabilities, so you might imagine that the country is also at the forefront of any interest in deep fake makers.
However, our rankings show that the country is decidedly uninterested in the creation of deep fakes when compared with other countries on this list. The US ranks joint second alongside the Netherlands and Singapore for searches for deep fake nude makers, but is middling in the ranking for other related terms.
Nevertheless, many high-profile US celebrities have already been the targets of deep fake videos and images posted online – and the US is already in the process of drafting new laws to protect individuals against the harm deep fake content can cause.
Despite being considered the second leading country for artificial intelligence, behind only the US, China emerges from our research as the least interested country in deep fake makers across all search queries.
China sits at the bottom of the ranking for searches for deep fake makers, including pornographic searches, suggesting that the capabilities of deep fakes may not have reached the country’s mainstream yet.
However, it’s also important to note that search data may not be 100% accurate for China-specific searches. AI is certainly not new to China – from deep fakes of livestreamers running 24/7 making headlines in 2023, to a rise in financial scams using deep fake technology. Perhaps more likely is the assumption that we are not able to access completely accurate search data from China on artificial intelligence-related queries.
To discern the authenticity of digital content and identify deep fakes, it’s important to know what tell-tale signs to look out for. The following indicators can help you differentiate between genuine content and sophisticated forgeries:
Deepfakes often struggle to replicate the nuances of natural lighting and shadows accurately. Look for irregularities in how light interacts with the subject, such as mismatched shadows or inconsistent lighting across different parts of the scene.
Pay close attention to the subject of the video of imge’s facial expressions and features. Deepfakes will often exhibit anomalies like unnatural blinking, odd lip movements, or facial expressions that don’t align with the emotional tone of the voice or situation. As technology gets more advanced, these anomalies will be more and more slight.
In deep fakes, the synchronization between audio and visual elements can be off. Check if the subject’s lip movements are in perfect sync with the spoken words. Any delay or mismatches between the visual and audio could be a red flag that you’re looking at a deep fake.
Deepfakes can sometimes struggle to replicate human skin texture realistically. Look out for any signs of abnormal skin texture or shifting skin tones that don’t seem natural or consistent throughout the video.
Analyze the background for any signs of distortion or fluctuation. Deepfake technology might focus on the subject and overlook the background, leading to anomalies in the surrounding environment.
Be cautious of videos that are presented as official, but intentionally presented in low resolution or quality. Sometimes, creators of deep fakes may use lower quality to conceal imperfections and anomalies that would be more noticeable in high-definition.
To get these findings, our DailyAI team sourced the top 10 countries deemed to be at the forefront of AI in 2023. We then found the average monthly search volume for a number of deep fake maker-related terms, and found the average number of searches per 100,000 people to off-set any bias from countries with a larger population size.
The adoption of AI has caused an increased need for proper data governance, and companies…
OpenAI on Monday launched a new family of models called GPT-4.1. Yes, “4.1” — as…
Evan Brown serves as the Executive Director of EDGE (Economic Development Growth and Expansion) at…
All of the 400 exposed AI systems found by UpGuard have one thing in common:…
In the 90s, we collected Pokémon cards, in the 2000s, we all had a weird…
Medicaid has become a central point of a heated political battle, as Republican lawmakers push…
This website uses cookies.