Bucharest: Artificial intelligence–generated misinformation is sharply increasing the risk of digital fraud in Romania, costing consumers hundreds of millions of euros each year and eroding trust in online transactions, according to a new study released by Visa.
The research estimates that online scams now cause economic losses of more than 800 million Romanian lei (about €160 million) annually in Romania. The findings suggest that the growing sophistication of AI-generated content — from manipulated images and videos to convincingly written messages — has made it significantly harder for users to distinguish between legitimate information and fraud.
Romanians who are unable to tell apart AI-generated fake content from authentic material are more than five times as likely to fall victim to digital scams, the study found. Among those who struggle to identify manipulated content, 71 percent reported exposure to fraud, compared with 13 percent among users who said they could reliably recognise it.
Financial Losses and Emotional Toll
Victims of online scams in Romania lose an average of 731 lei per incident, the study found. But the consequences often extend beyond immediate financial damage. One in four victims said they spent more than 24 hours attempting to resolve the fallout from a scam, dealing with banks, platforms and authorities.
The emotional and behavioural impact is also pronounced. More than a third of Romanians who have been targeted by digital scams said they became hesitant to make online purchases afterward, citing anxiety, stress and diminished trust in digital services.
“These are not just monetary losses,” said Elena Ungureanu, country manager for Visa Romania. “They involve lost time, emotional strain and a broader erosion of confidence in online channels.”
AI as Both Tool and Threat
While artificial intelligence has become a cornerstone of modern digital services — improving personalization, efficiency and fraud detection — it is also being rapidly adopted by criminal networks, the study found. Fraudsters increasingly use AI to craft persuasive messages, clone voices, generate fake endorsements and manipulate images, blurring the line between real and fabricated content.
“It has become easier than ever to confuse what is fake with what is real,” Ms. Ungureanu said, adding that Visa is investing heavily in AI-driven security tools in collaboration with industry partners to counter these risks.
Behavioral Vulnerabilities Online
The study highlights how everyday online habits are amplifying exposure to fraud. Users who share or post digital content without verifying its accuracy were significantly more likely to fall victim to scams — 37 percent, compared with 30 percent among those who check information before sharing.
Nearly one in three Romanians said they had shared online content at least occasionally without confirming its authenticity, a pattern that researchers say creates fertile ground for misinformation-based fraud.
Quickly scanning headlines, impulsive engagement with AI-generated content and redistributing posts without scrutiny have become common behaviors — and, increasingly, exploitable weaknesses.
Demand for Safer Digital Payments
Despite rising fraud risks, the study found strong demand for secure digital payment solutions. Nearly 80 percent of Romanian consumers said security was the most important factor when choosing an online payment method, and 90 percent said they would opt for a service offering stronger protections, even if it meant sacrificing some convenience.
The research was conducted by Opinium on behalf of Visa between August and September 2025, surveying 9,500 adultsacross 11 European countries. In Romania, the findings were based on a nationally representative sample of 1,000 adults.
As AI-driven content becomes more pervasive, the study suggests that improving digital literacy — alongside stronger technological safeguards — may be critical in preventing fraud and preserving trust in the online economy.
