What are Southeast Asian countries doing to combat deepfake fraud?

Deepfakes are becoming a serious threat in Southeast Asia, causing significant damage to businesses, governments and individual users.

According to Mr. Tan Ah Tuan, Head of Research and Innovation (Ensign Labs) at cybersecurity company Ensign InfoSecurity (Singapore), this technology not only reduces trust in organizations but also creates opportunities for identity theft, fraud, sensitive data leakage and causes severe reputational damage.

In addition, deepfakes also cause complex social problems, blurring the line between reality and fiction, making information in cyberspace difficult to distinguish and easy to manipulate.

Deepfakes are synthetic images or videos created using artificial intelligence (AI) technology. According to the University of Virginia (USA), an advanced type of machine learning (ML) called "deep learning" is the tool behind the creation of deepfakes. This algorithm learns to recreate content based on examples provided by programmers.

What sets deepfakes apart from Photoshopped images or face swapping technology is their ability to create sophisticated fake images that are almost indistinguishable from real photos.

According to the Global Initiative against Transnational Organized Crime, high-tech crime groups in Southeast Asia have taken advantage of AI to impersonate celebrities, thereby carrying out acts of extortion, fraud and spreading misinformation, causing serious damage to the people.

The organization's report shows that the Asia-Pacific (APAC) region witnessed a spike of up to 1,530% in deepfake-related incidents during the period 2022 - 2023. Notably, the Philippines had the highest growth rate, reaching 4,500%, while Vietnam recorded the highest rate of increase in deepfake fraud cases, at 25.3%.

What Southeast Asian governments are doing to combat deepfakes

Deepfakes from videos, images to audio recordings have become popular thanks to the great advances in Large Language Models (LLMs) like MidJourney, ChatGPT from OpenAI and Gemini from Google.

At the World Economic Forum on Cybersecurity 2023, 46% of survey respondents expressed concerns about the "confrontational" potential of AI, including threats such as fraud, malware development, and deepfake creation.

Here are some of the solutions that the Association of Southeast Asian Nations (ASEAN) is implementing to protect the community from the growing problem of deepfake fraud.

Develop policies and regulations

Regional governments are actively considering the development of policies and regulations related to personal data protection, privacy and user consent. The biggest challenge is how to promote innovation and enable businesses to achieve profits while ensuring that negative impacts on the public are minimized.

Singapore, for example, is ramping up cybersecurity to combat deepfake scams to protect its democracy and the country’s future. In one high-profile case, cybercriminals used the image of former Prime Minister Lee Hsien Loong to promote fake investment products.

In addition, leading technology corporations such as Microsoft, Meta, X (formerly Twitter), and Google also actively participate in management, and announced a joint commitment in February last year to combat the use of AI for fraudulent purposes.

Establishenforcement mechanisms

Researcher Carol Soon, head of the Institute of Policy Studies at the National University of Singapore, stressed that technology companies need to adopt a multi-pronged approach in deploying and enforcing measures to deal with the misuse of AI for fraudulent purposes.

ASEAN governments have been proactive in enacting laws to combat deepfakes and other online threats. Singapore, for example, has passed the Online Harm Act (OCHA) and the Cybersecurity Act to strengthen its cyber defenses.

Meanwhile, the Philippines established the National Computer Emergency Response Agency to handle digital incidents, while Indonesia launched the National Cyber ​​Security and Cryptocurrency Agency (BSSN) to deal with cyber-related risks.

Developing solutionscreative to fightdeepfake again

Research by UK technology company Sumsub shows that the number of deepfakes detected globally has increased 10-fold across many industries from 2022 to 2023.

The online media industry saw the sharpest increase in identity fraud. In particular, the cryptocurrency industry became the main target, accounting for 88% of detected cases in 2023, followed by the financial technology (fintech) industry at 8%.

Cybersecurity companies like Singapore-based GROUP8 are offering AI-powered solutions to detect and prevent cyberthreats. Startups should consider developing new technologies to detect and authenticate deepfakes, as well as training LLMs to accurately identify deepfakes in real-time scenarios.

Some possible solutions could include the use of digital watermarks, blockchain technology, and metadata to detect AI-generated content.

Advanced technology training

According to the World Economic Forum (WEF), raising public awareness of online dangers will help citizens become more vigilant and learn to distinguish between authentic information and fabricated content. This will also help prevent the spread of AI-generated misinformation.

Some helpful tips for identifying deepfakes include checking for consistency between the facial features of the person in the video and previous images, looking at the details in the photo or video, and assessing whether the character’s movements appear natural or cartoonish. Additionally, implementing multi-factor authentication will help protect personal devices from unauthorized access.

Prospectson combating deepfakes in Southeast Asia

As ASEAN pushes ahead with its digital transformation goals, the region is becoming vulnerable to cybersecurity threats. To combat the situation, companies need to deploy a range of security tools, including cloud technology, blockchain, automated processes and other technical solutions to ensure data safety and security.

However, the fight against deepfakes still faces many major challenges. Issues such as limited resources, the constant development of technology, a shortage of AI expertise, increasingly realistic deepfakes, inconsistent legal regulations, and low detection rates are significant obstacles.

To effectively deal with this threat, regional governments and stakeholders need to strengthen cooperation with each other, and seek input from other countries on the measures they are implementing.

In addition, providing adequate funding and resources is crucial to ensure anti-deepfake efforts are implemented effectively.