The Paradox of AI: How it Fuels and Fights Ecommerce Fraud
Q&A with Aviram Ganor, General Manager of EMEA, Riskified
How are eCommerce retailers being targeted by fraudsters, and why might they not even realise it?
Retailers are often targeted by fraudsters in subtle and complex ways, making it difficult for merchants to realize they’re being exploited. Fraudsters use online forums like Reddit and Telegram to share advice, guidance, and fraud-as-a-service kits aimed at exploiting merchants.
These communities thrive on the Dark Web, where participants can learn new techniques and access tools that help them exploit merchant vulnerabilities. The increasing use of generative AI, like WormGPT, has made it easier and faster for fraudsters to conduct these activities.
Are these online fraudster communities using AI to target retailers?
Yes. Online fraudster communities are using AI tools like WormGPT, the malicious counterpart to ChatGPT, to automate and scale fraudulent activities. These tools allow fraudsters to quickly generate bots that can exploit retailer systems, such as buying popular items the moment they drop.
Generative AI makes it easier for individuals, even with limited technical expertise, to carry out these scams. This has significantly lowered the barriers to entry, enabling more people to become involved in fraud.
Inversely, are AI-based fraud detection models helping retailers protect themselves?
Yes again. AI-based fraud detection models have been transformative for merchants. Unlike traditional fraud detection methods that rely on human-set rules, AI models continuously evolve by analyzing transaction data and detecting patterns of suspicious behaviour in real- time.
These models don’t require strict rules beforehand; instead, they learn to identify anomalies in customer behaviour, helping retailers detect fraud before it happens. With real-time insights, merchants can adjust their responses to emerging threats, enhancing their protection.
What are some of the challenges merchants face when dealing with policy abuse like return fraud, and how is AI helping?
Policy abuse, such as return fraud (e.g., "wardrobing" or replacing items with substitutes), is a significant challenge for merchants, particularly as it’s often perpetrated by customers who have previously been "good" shoppers. Research shows that three in four online merchants feel overwhelmed by this issue, and 84% find it increasingly difficult to detect abuse of returns and refund policies. AI models help merchants run real-time risk assessments, enabling them to set more dynamic and flexible policies. For example, a loyal customer may be allowed flexible returns, while a customer with a suspicious history may face additional scrutiny or fees. This adaptability helps merchants handle policy abuse more effectively.
How can AI help merchants create more agile and dynamic return policies?
AI allows merchants to implement dynamic and agile return policies that are tailored to individual customer behavior. By continuously assessing risk in real-time, AI helps retailers adjust their return policies based on the customer's history. For example, a loyal customer with a positive track record might be offered flexible returns, while a customer who has a history of return fraud could be required to pay a fee or face stricter conditions. This helps merchants minimise fraud while maintaining a positive experience for trustworthy customers
How does AI empower retailers to fight back against fraudsters during peak shopping seasons?
AI empowers retailers to stay ahead of fraudsters by providing real-time insights and proactive fraud detection capabilities. With the rise of emerging technologies used by fraudsters, retailers need to remain vigilant. By embracing AI, merchants can identify fraudulent transactions before they cause significant damage, even during high-risk seasons like the holidays. This proactive approach helps protect profits while ensuring loyal customers aren't impacted by overly strict fraud prevention measures.
