Artificial intelligence has fundamentally changed the landscape of online fraud, stripping away the obvious markers that once made scams easy to identify. Where typo-filled emails, poorly constructed websites and suspicious accents once served as warning signs, criminals today wield sophisticated AI tools to create nearly flawless impersonations that fool even the cautious. This technological shift means that traditional defensive strategies no longer suffice—and Malaysian and Southeast Asian consumers, increasingly active on social media and e-commerce platforms, face mounting risks from attacks tailored to their interests and vulnerabilities.
The scale of this problem is staggering. The FBI reported that Americans lost nearly US$21 billion (RM85.51 billion) to cybercriminals last year, with approximately US$893 million (RM3.63 billion) of those losses directly attributable to AI-enhanced fraud schemes. These figures represent just one market; the global toll is substantially higher. As digital adoption accelerates across Southeast Asia, with millions of new users coming online in Indonesia, the Philippines, Vietnam and Thailand, the region is becoming an attractive target for sophisticated criminal networks that exploit language barriers and varying levels of digital literacy.
One particularly pernicious form of AI fraud involves the creation of counterfeit e-commerce sites that mirror legitimate brands. A researcher nearly fell victim to such a scheme after encountering a TikTok advertisement offering Hoka sneakers at an 80 percent discount. The website that loaded appeared authentic, complete with professional design and convincing product listings. However, a simple Reddit search revealed that numerous users had been defrauded by the same operation, and the shoe brand had publicly warned consumers about a wave of fake outlets using its name. This type of scam proves especially effective on social media platforms popular in Malaysia and across Southeast Asia, where users may be less familiar with the subtle differences between genuine and fraudulent sites.
Social media platforms have become primary vectors for these scams, prompting legal action and regulatory scrutiny. The Consumer Federation of America filed a complaint against Meta, alleging that the technology company has misrepresented its efforts to combat fraud on Facebook and Instagram. Additional legal complaints from Santa Clara County in California highlight the scope of the problem. In response, Meta disclosed that it removed 159 million scam advertisements and suspended nearly 11 million accounts associated with known fraud operations during the previous year. The company has committed to deploying new technologies to identify and remove such content more rapidly. TikTok similarly maintains that it prohibits deceptive advertising practices, reporting that in the final quarter of 2025, it removed 97 percent of violating spam content before users reported it.
AI has made it possible for fraudsters to craft attacks of unprecedented personalization, moving beyond the generic phishing attempts that once characterized internet crime. Rather than contacting strangers, scammers now impersonate family members, romantic partners and trusted acquaintances—exploiting the personal information readily available on social media and other public databases. Andrew Yoon, a researcher at CivAI, an organisation dedicated to educating the public about artificial intelligence capabilities, explains that creating convincing real-time video calls with realistic body replacement and voice synthesis has become extraordinarily cheap and simple. This capability represents a fundamental threat to elderly relatives and vulnerable populations who may struggle to distinguish deepfakes from reality.
The romance scam has evolved into a particularly insidious version of fraud enabled by AI video synthesis. A lonely individual might receive a message claiming to be from someone from their past, eventually leading to a video call with a convincingly rendered AI simulation of that person. The technology allows criminals to target victims based on psychological profiles and personal weaknesses. Job seekers have been deceived by fake AI interviewers into performing unpaid work for non-existent companies, losing both time and money. Given that phone numbers can be easily spoofed and personal details about family members are publicly accessible online, such schemes can achieve a disturbing level of authenticity that bypasses natural skepticism.
Yoon recommends a decidedly low-technology solution: families should have explicit conversations about the possibility of impersonation scams, particularly with older relatives less experienced with digital technology. Establishing a pre-arranged code word or security phrase can serve as a verification mechanism when doubt arises during unexpected communications. This approach harks back to traditional security practices but remains surprisingly effective in an age of technological sophistication. For Malaysian households, where extended family networks often span multiple countries and communication across time zones is common, establishing such protocols could prevent costly fraud.
Celebrities and high-profile business figures have become unwitting tools in AI-powered scams. The proliferation of deepfake videos featuring Hollywood personalities and corporate executives has created an environment where anyone with internet access can be impersonated convincingly. Gordon Ramsay, the celebrity chef, experienced this firsthand when deepfake videos of him circulated across social media promoting nonexistent cookware giveaways. Victims who believed they were paying nominal shipping fees for free frying pans were instead handing their credit card information to criminals. Richard Branson, founder of Virgin Group, faced similar exploitation when fraudsters used deepfake videos to encourage his admirers to make fake investments. Branson eventually posted an educational video on Instagram warning his followers to verify all communications through official company channels before acting on celebrity endorsements.
The sophistication of AI scam advertising exploits the same targeting mechanisms that legitimate businesses use to reach consumers. Scammers purchase advertising space on TikTok and Instagram, using the platforms' targeting tools to identify people interested in specific products—such as bicycles, athletic gear or electronics. What differentiates these operators from legitimate advertisers is their business model: they have no actual products to deliver, making their advertising costs negligible compared to the revenue extracted from defrauded victims. This economic calculus makes such campaigns highly profitable despite relatively low conversion rates. For Malaysian consumers browsing these platforms during their commute or leisure time, distinguishing between genuine advertisements and sophisticated scams has become considerably more difficult.
Consumers can employ several verification strategies to protect themselves from fake e-commerce sites. The simplest approach involves searching the website's URL on Google and reviewing community feedback on platforms like Reddit, where victims frequently document their experiences. A more comprehensive method involves using AI itself as a defensive tool. Malwarebytes recently partnered with OpenAI and Anthropic to integrate scam-detection capabilities into ChatGPT and Claude. Users can paste website addresses and screenshots into these chatbots and request analysis of whether a site is legitimate. This application of AI to counter AI-enabled fraud represents a practical adaptation to the changed threat landscape, though it requires effort and technical familiarity that not all consumers possess.
Beyond these specific tactics, one timeless principle retains absolute validity in the age of artificial intelligence: offers that seem unrealistically generous warrant immediate skepticism. An 80 percent discount on premium sneakers, a job offering substantial pay for minimal qualifications, or an unexpected inheritance from a distant relative should trigger careful verification before any personal or financial information is shared. This principle transcends technological advancement because it reflects fundamental economic logic—legitimate businesses cannot sustain operations through unsustainably generous offers. For Malaysian consumers navigating the increasingly complex digital marketplace, maintaining this simple heuristic may prove the most reliable defense against becoming another statistic in the expanding catalogue of AI-enabled fraud.
