Artificial intelligence (AI) is making tremendous strides, opening up unprecedented opportunities for economic development, innovation, and improved quality of life. However, along with these enormous benefits, AI is also becoming a tool exploited to carry out increasingly sophisticated, large-scale, and hard-to-detect fraud, urgently requiring the international community to prevent the risk of AI being turned into a "weapon" by criminals.
When AI is misused for illicit purposes, the consequences can spread at an unprecedented rate. Therefore, the global tightening of AI regulation is not aimed at stifling innovation, but at ensuring that technology develops safely, responsibly, and in the common good of society. Digital fraud is becoming increasingly sophisticated.
The development of AI is bringing with it negative consequences, including the potential for it to be exploited as a tool to aid fraudsters.
The explosion of AI is fundamentally changing the way cybercriminals operate. While previously fraudsters relied primarily on spam messages, fake emails, or simple scripted calls, AI has now automated almost the entire fraud process, from data collection and victim psychology analysis to creating highly convincing fabricated scenarios.
One of the most alarming technologies today is deepfake, which uses AI to create fake images, videos , or voices that are so realistic they are difficult to distinguish with the naked eye. With just a few seconds of voice or image data collected from social media, malicious actors can create fake calls or videos impersonating relatives, business leaders, business partners, or even government officials.
According to the latest report from the International Criminal Police Organization (Interpol), cybercrime now accounts for more than 30% of all recorded crimes in over half of the countries surveyed in the Asia-Pacific region. In 2024, more than 6.5 billion cyber threats were detected and neutralized in the region.
Recently, a shocking incident occurred in Hong Kong (China) where a company was scammed out of $25 million after an online meeting where all the participants were actually products of deepfake technology. Even more concerning, AI not only helps to fake images and voices but also has the ability to interact in real time. AI systems can chat, answer questions, and adjust content based on the victim's reactions, making scam scenarios more convincing than ever.
In Malaysia, the number of online scams in 2025 is projected to reach 66,204, an 87% increase from 35,368 in 2024, with investment scams causing losses of up to 1.47 billion ringgit. Malaysian Inspector General of Police Mohd Khalid Ismail stated that scams are becoming increasingly sophisticated as criminal gangs exploit technological advancements and modern communication platforms.
In Thailand, the Securities and Exchange Commission (SEC) reported that in the first five months of 2026, it handled eight criminal cases related to stock price manipulation, fraud, providing false information, and illegal trading in the capital market, with a total of 43 individuals accused of violations. In a crackdown on online investment scams, Thai authorities froze 368 accounts and trading channels that violated regulations. Processing times ranged from 7 to 48 hours from the time information was received, helping to limit losses for the public.
In Europe, according to data from the Belgian Financial Services and Markets Authority (FSMA), in the last six months of 2025 alone, Belgian consumers reported total losses of over €23.4 million due to investment fraud. Throughout 2025, the FSMA received 2,911 complaints related to illegal activities, an increase of approximately 11% compared to 2024. Fraudulent online platforms linked to cryptocurrencies accounted for the majority of these losses…
Warning about the risks
In response to AI-powered scams, leading security agencies belonging to the Five Eyes intelligence alliance—comprising the US, UK, Australia, Canada, and New Zealand—issued a strong warning on June 23, 2026, about the increasing risks posed by artificial intelligence (AI), stating that the world needs to act urgently before this technology surpasses human control.
In a joint statement, the Five Eyes security alliance noted that the pace of AI development is now outpacing the development of security management and protection mechanisms. Experts warn that AI is no longer simply a technology of the future but has become a tool exploited by cybercrime groups and hostile forces for dangerous purposes.
One of the biggest concerns is the increasing use of large model languages (LLMs) to automate the process of writing malicious code. This allows even those with limited technical skills to create sophisticated cyberattack software, from ransomware to intrusion tools capable of bypassing multiple layers of traditional security. In addition, deepfake technology (AI-generated images and voices) is also considered to be developing at an alarming rate. AI-generated fake images, videos, and voices have now achieved a very high level of realism, becoming tools for large-scale phishing campaigns, information theft, or disseminating misinformation. According to the Five Eyes alliance, the risks extend beyond cybercrime and financial fraud. AI can also be exploited to manipulate public opinion, spread fake news, and interfere with democratic processes, thereby directly impacting the national security of many countries.
In light of this reality, five countries have issued a series of urgent recommendations to mitigate the risks posed by AI. The focus of these proposals is a shift from a "post-incident" approach to a "proactive prevention" model. The alliance calls for major technology corporations to take greater responsibility for the AI products they develop. Accordingly, cybersecurity measures should be integrated from the system design stage, rather than only being added after the product is in use. Security agencies also emphasize the importance of increased real-time data and intelligence sharing between the public and private sectors to detect AI-related security vulnerabilities early. Countries are tightening AI regulation to prevent a wave of technology scams .
Faced with the increasingly serious problem of AI-powered fraud, many countries and international organizations have begun implementing strong regulatory measures to curb the misuse of AI. The European Union (EU) is leading this effort with the implementation of the AI Act, considered the world's first comprehensive law on the regulation of artificial intelligence.
According to the EU's AI Act, published on June 10, 2026, and effective August 2, 2026, providers of chatbots and AI systems that interact directly with users must clearly state the non-human nature of such interactions. Providers of systems that generate aggregated content from images, audio, video, and text must ensure that it is marked up in a machine-readable format. Deepfakes and AI content in public affairs must be clearly labeled for end users. Penalties for violations can reach €15 million or 3% of the organization's annual global revenue, whichever is higher. In some cases involving prohibited acts, penalties can reach €35 million or 7% of revenue.
In addition to the AI Act, the European Commission also published a Code of Practice on AI Content Transparency, guiding businesses to adopt measures such as digital watermarks, metadata, and unified AI identifiers across the bloc. At the national level, many countries such as France, Germany, and Italy are amending their criminal laws to directly address the use of deepfakes for fraudulent, defamatory, or misappropriation purposes.
Meanwhile, financial regulators in Belgium are continuously updating their lists of fraudulent investment platforms and intensifying warnings to the public. However, experts agree that while the law is important, it cannot be the only solution.
Phuong Vy
