Deepfakes in Southeast Asia are causing significant damage to enterprises, governments and individual users. According to Tan Ah Tuan, Head of Ensign Labs at Ensign InfoSecurity, they erode trust in institutions, enabling identity theft and fraud, leaking sensitive data, and making people suffer reputational losses. Moreover, they are creating societal issues by blurring the lines between reality and fiction in the informational space.

A deepfake is an artificial image or video created through artificial intelligence (AI). The University of Virginia says a type of machine learning (ML) known as “deep learning”, whereby an algorithm learns how to create specific content based on the examples the programmers gave it, creates deepfakes. They differ from photoshopped or face-swapped images because the technology produces fake photos that are indistinguishable from real ones.

Tan Ah Tuan, Head of Ensign Labs discusses navigating digital deception with deepfake detection

For example, the Global Initiative Against Transnational Organised Crime noted that tech-savvy criminals in Southeast Asia were using AI to impersonate public figures and hurt citizens through extortion, fraud, and misleading information. According to the organisation, the Asia-Pacific (APAC) had a 1,530% increase in deepfake cases between 2022 and 2023. The Philippines had the highest growth at 4,500%, and Vietnam had the highest increase in fraud at 25.3%. 

Steps Southeast Asian governments are taking to fight deepfakes

Deepfakes of videos, photos, and audio recordings have become widespread due to advancements in Large Language Models (LLMs) like MidJourney, ChatGPT from OpenAI, and Gemini from Google. At a World Economic Forum on Cybersecurity in 2023, 46% of respondents said they were concerned about the “adversarial capabilities” of AI, such as phishing, malware development, and deepfakes.

Here are a few steps the Association of Southeast Asian Nations (ASEAN) is taking to protect the public:

Policies and regulations

Regional governments are analysing the policies and regulations about personal data protection, privacy, and user consent. The challenge is to foster innovation and enable companies to achieve profitability while preventing harmful impacts on the public.

As an example, Singapore is addressing the deepfake scams by increasing security to protect democracy and safeguard the country’s future. In one incident, a cybercriminal used former Prime Minister Lee Hsien Loong’s likeness to promote investment products.  

Leading tech companies, including Microsoft, Meta, X, and Google, are also playing a role in regulation, announcing a joint commitment in February to combat the deceptive use of AI in elections this year. 

Establishing enforcement mechanisms

Carol Soon, principal research fellow and head of the Institute of Policy Studies in Singapore’s Society and Culture department, said a multi-prong approach is needed for tech companies to implement and enforce the fight against deceptive AI use.

ASEAN governments have implemented laws to combat deepfakes and other online threats. For example, Singapore passed the Online Criminal Harms Act (OCHA) and the Cybersecurity Act. The Philippines has the National Computer Emergency Response to deal with digital incidents, whereas Indonesia has the National Cyber and Crypto Agency (BSSN).

Innovating anti-deepfake solutions

Sumsub Research highlighted a 10X increase worldwide in deepfakes detected in multiple industries between 2022 and 2023. The online media industry experienced the highest rise in identity fraud cases. The Crypto sector was the main one targeted, representing 88% of cases detected in 2023, followed by the Financial Technology (fintech) sector at 8%.

Cybersecurity companies like Singapore-based GROUP8 offer AI-based solutions to detect and prevent online threats. Startups should consider creating multiple detection and authentication technologies and train the LLMs to accurately identify deepfakes in real-time situations. They can spot AI-generated content using digital watermarks, blockchain, and metadata.

Education/tech literacy

According to the World Economic Forum (WEF), increasing public awareness of online risks teaches citizens to be vigilant and to learn to differentiate between factual and fabricated content. It also prevents people from sharing false AI-generated content.

Some tips for spotting deepfakes include checking whether a person’s features are consistent with past images, fact-checking the details of the photo or video, and assessing whether the individual’s movement looks realistic or cartoonish. Additionally, multi-factor authentication prevents unauthorised access to personal devices.

Outlook on combatting deepfakes in Southeast Asia

Projections indicate that the Cybersecurity market will reach USD 4.37 billion in 2024 due to being aware of data risks and threats. With ASEAN pursuing its digital transformation objectives, the region is vulnerable. Companies must deploy various tools, cloud and blockchain technologies, and automated processes, among others, to handle the situation.

Challenges that continue to affect the deepfake fight include limited resources, constantly evolving tech, inadequate AI expertise, hyper-realistic deepfakes, inconsistent regulations, and low detection rates.

Nevertheless, the quality of deepfakes in Southeast Asia will likely improve as technology advances. Therefore, regional governments and other stakeholders need to collaborate, seek input from foreign countries on the steps they are taking, and provide adequate funding and resources to combat the threat.