Mitigating APP fraud through behavioural economics

With mandatory reimbursement rules now in force, sending and receiving payment service providers (PSPs) share liability for APP fraud losses up to £85,000 – resulting in £112 million reimbursed in the first nine months. Fraudsters are innovating faster than ever, using AI-driven impersonation and social engineering to bypass traditional controls. Deepfake video calls and voice cloning have already enabled multimillion-pound scams globally, and UK businesses report a sharp rise in AI-related fraud attempts. 

Against this backdrop, firms must not only strengthen technical controls but also meet their obligations under Consumer Duty, preventing foreseeable harm and supporting good outcomes by designing interventions that reflect real human behaviour. 

Why behavioural economics matters

APP fraud remains one of the most damaging forms of financial crime in the UK. In 2024, losses exceeded £450 million across 185,000 cases, according to UK Finance. Victims are tricked into authorising payments, often through convincing impersonation or promises of quick rewards.  

Research identifies four key behavioural biases that make consumers vulnerable:  

Scarcity  

Scarcity diminishes cognitive resources, causing people to focus intensely on unmet needs and make poor decisions. Economic scarcity is a clear example of this where fans, desperate for concert tickets, often fall for fake offers at discounted prices. Similarly, emotional scarcity (such as loneliness) can make individuals more susceptible to romance scams.

Willingness to trust  

Faced with constant decisions and competing demands, people rely on educated guesses to simplify choices. For example, trust in familiar brands or institutions reduces scrutiny, which fraudsters exploit by mimicking the branding and tone of banks or government agencies.

Similarity bias

People judge plausibility based on similarity to what they perceive as typical. Fraudsters exploit this by replicating real transactions with professional graphics, product descriptions and fake reviews, making scams appear legitimate.

Fast system 1 thinking  

Leading behavioural economists often draw a distinction between fast, intuitive ‘System 1’ thinking and slow, deliberative ‘System 2’ thinking. APP fraudsters deliberately create urgency or panic to trigger System 1 responses, bypassing rational checks and balances that are present in more nuanced deliberative System 2 thinking. 

Purchase scams often combine all four:  

• offering products at unusually low prices (scarcity)
• impersonating trusted brands (trust)
• using realistic product details (representativeness)  
• insisting the deal is time-limited (System 1).  

These biases rarely operate in isolation. In fact, most scams combine several of them to maximise impact.

What the PSR paper reveals

The PSR’s research maps behavioural biases to specific scam types – from investment fraud to impersonation – and explores interventions that could make a real difference. Recurring themes include:

• introducing friction at the right moment, such as deliberate pauses or extra verification steps when risk signals are high
• deploying contextual warnings tailored to the transaction type, using language that resonates emotionally rather than generic alerts
• designing choice architecture so that the safest option is the easiest, reducing reliance on rushed decisions
• using social proof and feedback loops to show consumers how many scams have been stopped or how others avoided fraud, reinforcing caution. 

The report also warns against overuse of warnings. Frequent alerts can lead to ‘alert fatigue’ reducing their effectiveness when a genuine scam occurs. Interventions must be tested rigorously, ideally through randomised controlled trials – a method widely used in behavioural economics to measure impact.

Balancing risk and trust: The evaluation tightrope in APP fraud policies

Read this article

Turning insight into action  

When applied well, behavioural economics can transform fraud prevention strategies. Embedding behavioural nudges into payment flows can reduce fraud without degrading user experience. Tracking consumer responses can feed predictive analytics and improve detection. And demonstrating proactive, evidence-based protection builds trust with customers and regulators alike.

This is about shifting from reactive reimbursement to proactive prevention and doing so in a way that reflects how people actually behave. 

Three areas for firms to work on

To support customers and prevent foreseeable harm, PSPs can embrace behavioural economics and structure their efforts around the following three core stages.

1 Diagnose vulnerabilities

Start with a behavioural audit. Map where payment processes intersect with high-risk behaviours, such as urgency, trust cues and decision fatigue. This diagnostic step helps identify the points where fraudsters are most likely to exploit cognitive biases.

2 Design targeted interventions

Once vulnerabilities are clear, develop interventions that counteract these biases. For example:

• if scams exploit scarcity, introduce prompts that trigger reflection: “Is this too good to be true?”
• if trust is being misused, embed checks that encourage scrutiny of identity and credentials
• if urgency drives poor decisions, add friction such as timed delays or multistep confirmations to slow the process and engage slower System 2 thinking.

Testing is critical. Use randomised controlled trials to measure effectiveness and avoid unintended consequences such as alert fatigue or false reassurance.

3 Deliver and refine continuously

Firms need to roll out interventions in a way that balances fraud prevention with customer experience. This includes collaboration across banks, PSPs and technology platforms to avoid a fragmented approach and alert fatigue. When doing so, it’s important to regularly monitor outcomes as behavioural responses evolve over time. Strategies that work today may lose impact tomorrow, so firms need to adapt dynamically.

Practical measures include embedding social norms into customer messaging, for example by stating that ‘Most customers take a moment to verify before sending money’. Firms could also introduce intelligent friction for high-risk transfers or leverage AI-driven behavioural analytics to detect anomalies such as erratic navigation patterns or unusual transaction timing. These steps provide an additional layer of protection without compromising usability. 

The bigger picture 

APP fraud is a human problem as much as a payments problem. By recognising the psychological levers used during APP fraud, firms can design defences that work with human behaviour rather than against it. This will help firms reduce APP fraud, helping to prevent foreseeable harm and promote good outcomes under Consumer Duty.  

For more information on behavioural economics and fraud prevention, contact Tom Middleton and Alison Kopra.