AI-Powered Fraud: How Image Generation Threatens Restaurants

An upgraded AI function can now generate incredibly realistic images, including fake restaurant receipts. This new capability opens the door for unscrupulous individuals to try and scam restaurants out of free meals, gift cards, and more. To help you understand this emerging threat and learn how operators can protect their bottom line, Modern Restaurant Management (MRM) magazine consulted Doriel Abrahams, Principal Technologist at Forter, a leader in e-commerce trust and fraud prevention. His insights are crucial for navigating this new wave of AI-powered scams. With more than a decade of hands-on experience using AI to monitor emerging fraud trends, attacker MOs and advanced technologies, Abrahams brings a unique perspective to addressing merchants' most complex challenges.

What are some things restaurant operators do to protect their businesses from AI-generated fraud?

The use cases for AI-generated fraud are growing by the day. ChatGPT’s generated image capabilities are particularly concerning for the food & beverage industry because abusers may try to complain about poor service or food quality using fake receipts to get freebies, gift cards or comped meals. Supplier fraud could also be an issue with bad actors using fake invoices to confuse restaurant or franchise owners.

The use cases for AI-generated fraud are growing by the day.

For this reason, businesses should remain vigilant and take extra precautions if a transaction or complaint seems remotely suspicious. Cross-validating receipt numbers against internal systems and records can provide much-needed context. While GenAI is good enough to make a "believable" receipt, it will never have the right ID number. 

Similarly, if businesses receive an unexpected invoice from a supplier, the best way to ensure its validity is to take a few minutes to call and confirm the document with the business’s trusted and known contact. While some may think the time lost is not worth the extra step, the cost of fraud is much greater.

How should operators better inform themselves about AI concerns for their business?

In addition to maintaining a healthy fraud prevention and management posture for their business, restaurant owners should stay up to date with the latest AI developments, especially as it relates to fraud tactics within the food and beverage industry. Keeping an eye on emerging trends and headlines surrounding AI acceleration will help restaurant owners and employees stay aware of how bad actors may attempt to commit fraud.

How should operators train their staff to look for signs of fraud?

In recent years, fraud has become more refined and is conducted at greater scale – partly accelerated by AI. In the early days of digital commerce, food and beverage industry fraud included activities like card testing, account take overs, chargebacks, fake account creation and discount abuse. 

Today, GenAI continues to be a fantastic fraud accelerator. ChatGPT's image generator is just the latest proof point. It's no surprise that technology designed to help everyday users work faster also has very applicable use cases for bad actors looking to make their schemes more legitimate and convincing.

It’s crucial for businesses to continue to keep their employees informed on the latest developments of fraud in the food and beverage industry. It will enable them to spot suspicious activity and fraud attempts from fraudsters trying to take advantage of new technology like the image generator. Holding staff meetings and informative training sessions is a great way to introduce the topic to employees. Additionally, sharing real-world examples and use cases of recent AI-powered fraud in the industry would also be helpful for understanding what fraud could look like.

Many restaurants have the “Customer is Always Right” as a mantra. How can operators balance wanting to be responsive to guest complaints with not being taken advantage of by those who try to abuse the policy?

When it comes to fulfilling a restaurant order, time is of the essence. Customers expect orders and food service to be completed within a reasonable timeframe (think 30-90 minutes). This limited time can pose a strain on restaurants hoping to verify the validity of a customer complaint or order. 

It’s vital for restaurant operators who are vetting a customer complaint — while also ensuring their business is not being taken advantage of — to have a robust fraud protection system that will help keep peace of mind when confirming the identity of a customer and the order in question. Proactively communicating an estimated response timeframe will give your customer visibility into the process and provide your business time to dig into the case.