According to a new report from fraud prevention platform Ravelin, e-commerce merchants are bracing themselves for a significant rise in AI-powered fraud over the next 12 months. The majority (91%) of retailers, travel companies, digital goods providers, and marketplaces are worried about criminals turning to AI to attack them and their customers. There is concern that many different types of fraud will spike in the coming year.
Ravelin’s AI V AI report shows that 64% of merchants fear they have already been targeted by fraudsters using artificial intelligence. There is also concern that AI-enabled fraud will impact company growth (51%) and revenue (39%).
Fraudsters leverage AI-powered tools such as large language models (LLMs) to craft compelling phishing attacks, synthetic IDs to create fraudulent accounts at scale and voice cloning technology to deceive both consumers and businesses. These automated and scalable fraud techniques are making it more difficult than ever for merchants to distinguish between genuine customers and bad actors.
However, while the threat is real, Ravelin’s CEO Martin Sweeney says it’s important for the issue not to be exaggerated or misunderstood.
“I think the real story is that AI is part of a package of tools that help people become more efficient in perpetrating the crimes they already wanted to perpetrate,” Sweeney said. “There is no magic ‘help me be a really good fraudster’ button. But there is a series of improvements fraudsters can make to their process to make it harder to detect. Just as there have always been.”
“On our end, as fraud prevention experts, we want to do the same: Continuous improvements to our processes, strategies, and tooling to make ourselves more efficient and to be able to use at scale.”
Sweeney said that even if fraud teams have irrefutable proof that AI tools are used against their fraud department, this can be a simple, obvious application. For example, a criminal might use ChatGPT to write a refund abuse claim or a phishing email.
“Is that groundbreaking or impossible to prevent?” Sweeney asked. “I don’t think so.”
AI-powered fraud prevention tools are an effective line of defense, with machine learning and behavioral analysis emerging as critical weapons in merchants’ anti-fraud arsenals. Almost three-quarters (71%) of the companies spoken to are using machine learning (ML), large language models (LLMs) or any other AI technology to detect and/or prevent fraud.
Merchants believe AI-enabled fraud will have the most significant impact on online payment fraud (53%), fraudulent chargebacks (46%), and account takeover (ATO) attacks (42%). With AI-powered fraud techniques evolving rapidly, businesses must take proactive measures to protect themselves. The Ravelin report underscores the need for a multi-layered approach that combines AI-driven fraud prevention tools with human expertise.
Key features of Ravelin’s AI solutions include:
Bespoke machine learning models: Developed using each merchant’s unique data, KPIs, and goals, these models identify and block fraud in real-time.
Automated fraud prevention: AI-driven automation streamlines fraud operations, reducing the burden on merchants while improving accuracy.
Multi-model approach: New models for each client are continuously developed to improve results further, taking into account hundreds of data points across several families of ML features.
Natural Language Processing: Another type of AI, NLP is used to better understand non-numerical data within context, increasing the accuracy of fraud detection.
LLM/AI fraud rule creator: For convenience and speed, fraud managers and analysts can describe the fraud rules they want to create or test in their own words, in any language, and the AI rule creator will create it for them.