Why financial inclusion is now an AI story

About the Center

(Header) About the Center

The digital transformation of European micro-businesses

Focus Areas

(Header) Focus Areas

Scaling AI for inclusive growth

Events

(Header) Events

Building trust in the age of artificial intelligence

Insights

(Header) Insights

(Footer) About the Center

(Footer) Focus Areas

(Footer) Events

(Footer) Insights

PostContent

To even have a fighting chance in the battle against financial scams and fraud, in an increasingly digital and AI-driven landscape, we need to understand the challenge from a behavioral perspective. The problem is not unique to one region - it is a global imperative. Financial fraud and scams are increasingly widespread in the U.S. According to a 2026 Bankrate survey, <a href="https://www.bankrate.com/credit-cards/news/financial-fraud-survey/#key-insights" rel="noopener noreferrer" target="_blank" class="">40% of adults reported</a> experiencing some form of financial fraud or scam in the previous 12 months, up from 34% a year earlier, while 73% say they have experienced financial fraud or a scam at some point in their lives. In Southeast Asia, a recent survey across six countries from the Tech for Good Institute notes that 60% of respondents have encountered a scam in the past year, with 38% overall indicating they lost money, valuables or personal information as a result. The financial impact can be devastating, particularly for those that are low-income and among the most vulnerable. As financial transactions become more digital and multiply, rising at an <a href="https://www.oecd.org/en/publications/protecting-consumers-from-financial-scams-and-frauds_d41817bb-en/full-report/component-4.html#chapter-d1e200-f641103ea8" rel="noopener noreferrer" target="_blank" class="">estimated rate of 7.6% per year</a>, the risks to financial security simultaneously increase. The evolution of AI tools complicates the forecast, adding to the complexity of strategies used by fraudsters. In <a href="https://latam.kaspersky.com/blog/panorama-amenazas-latam-2023/26586/?srsltid=AfmBOooiNOKWCb0iRxJv5joDyQ8UcivjFpCiG4NaNWQPUC1VYdO6Fvw3" rel="noopener noreferrer" target="_blank" class="">Latin America</a><a href="https://latam.kaspersky.com/blog/panorama-amenazas-latam-2023/26586/?srsltid=AfmBOooiNOKWCb0iRxJv5joDyQ8UcivjFpCiG4NaNWQPUC1VYdO6Fvw3" rel="noopener noreferrer" target="_blank" class="">,</a> for instance, the number of phishing attacks increased six-fold between 2022 and 2023, mostly driven by post-pandemic economic reactivation together with technological advancements. Out of almost 300 million phishing attempts identified by Kaspersky, 43% were directed at financial information. In line with <a href="https://www.mastercard.com/news/latin-america/es/sala-de-prensa/comunicados-de-prensa/pr-es/2025/noviembre/a-la-vanguardia-frente-a-las-amenazas-ciberneticas/" rel="noopener noreferrer" target="_blank" class="">Mastercard’s recent research</a>, this data shows that phishing attempts are the third most common tactic, behind social media and phone scams. Dealing with scams and fraud requires what the Tech for Good Institute calls a <a href="https://techforgoodinstitute.org/research/tfgi-reports/building-resilience-against-digitally-enabled-scams-and-fraud-in-southeast-asia-a-whole-of-society-approach/" rel="noopener noreferrer" target="_blank" class="">“Whole-of-Society Approach”</a>. We need to understand how people experience it, what systems are in place to prevent it, and what solutions can be developed to help people recuperate afterwards. Along these lines, an upcoming article by the <a href="https://www.centerforfinancialinclusion.org" rel="noopener noreferrer" target="_blank">Center for Financial Inclusion at Accion</a> will explore the role of insurance in protecting individuals and small businesses against financial fraud.Our behavioral lens complements these approaches by surfacing how even the most sophisticated fraud attempts leverage human psychology to achieve their purpose, and describing how solutions can be better designed for end users. The time is ripe to use the same levers to protect and empower people to secure their financial lives.

<h3 class="">The psychology of a scam</h3>A few months after moving to the U.S., Hans a behavioral scientist and one of the authors of this piece, received a call from someone at the Social Security Administration (SSA) claiming that his Social Security Number (SSN) had been used to commit a crime and that they needed to verify his identity so he might avoid facing any charges. He remembers how his heart started racing and breathing became shallow. The caller persistently demanded personal information - name, date of birth and SSN. For a brief moment, doubt crept in, and Hans suggested going to the SSA offices to sort the issue instead of giving information over the phone. The person on the other end of the line escalated, threatening that he could get arrested at their offices. His response, counterintuitively, confirmed Hans’ suspicion that it was a scam and he hung up the phone. All of it lasted only 2-3 minutes. Most people are not lucky enough to doubt the situation and stop the scam. Research shows that most scam episodes leverage three <a href="https://webarchive.nationalarchives.gov.uk/ukgwa/20140402132426/http://www.oft.gov.uk/shared_oft/reports/consumer_protection/oft1070.pdf" rel="noopener noreferrer" target="_blank" class="">similar behavioral levers</a>: <ul class=""><li class="">First, they present an imbalanced situation of high risk (or high reward), that requires simple actions. For example, receiving a large payout by entering a fake raffle, or sharing account details to prevent losing your money. Behavioral science <a href="https://link.springer.com/article/10.1007/BF00122574" rel="noopener noreferrer" target="_blank" class="">research</a> suggests that people may become more willing to take a higher risk to prevent a loss, than to gain a similar reward. </li><li class="">Second, they also create a sense of urgency, such that an action cannot be carefully considered. Both of these levers trigger us to rely on our automatic thinking, where we rely mostly on heuristics and environmental cues to make decisions. We override our thoughtful processing of information, in lieu of - what we mistakenly feel - are quick and effective decisions. <a href="https://www.sciencedirect.com/science/article/pii/S0301051196052234?via%3Dihub" rel="noopener noreferrer" target="_blank" class="">Cognitive research</a> shows that in high pressure situations, we can enter a cognitive tunnel, where our attention solely focuses on a specific element (the scam) and we stop paying attention to things around it, often neglecting information that could help us make a better decision. </li><li class="">Finally, scammers count on this cognitive tunnel and the higher error rate of automatic decisions to pull a third critical lever: <a href="https://webarchive.nationalarchives.gov.uk/ukgwa/20140402132426/http://www.oft.gov.uk/shared_oft/reports/consumer_protection/oft1070.pdf" rel="noopener noreferrer" target="_blank" class="">trust</a>. Phone scams, in particular, use knowledge about the victim to build a relationship, sometimes signaling authority (i.e., being a bank associate or a government official). Meanwhile, digital scams tend to rely on visual and textual cues that resemble official channels, such as creating websites that are easily mistaken or using signatures and logos that signal official channels. As victims process information automatically and in a rush, they fall back on available cues to decide whether to trust the presented information. Labeled “bias-induced gullibility”, <a href="https://link.springer.com/article/10.1007/s10611-024-10187-1" rel="noopener noreferrer" target="_blank" class="">research from Malaysia</a> and the <a href="https://www.emerald.com/rbf/article-abstract/17/5/769/1278536/Experiencing-investment-scams-in-the-Philippines?redirectedFrom=PDF" rel="noopener noreferrer" target="_blank" class="">Philippines</a> suggests that in the presence of high rewards or relevant trust signals (references to known organizations or experts), people are more likely to overestimate the trustworthiness of said cues and are more likely to fall for a scam. </li></ul>Victim characteristics may also play a role in their likelihood of falling for a financial scam. Contrary to common stereotypes, older adults in the United States <a href="https://papers.ssrn.com/sol3/papers.cfm?abstract_id=6396678&amp;__cf_chl_f_tk=RbR53Sac.RkyGfHgMS4ESEZRV5vSH7JY0jew7gashlo-1783370586-1.0.1.1-igPTvlP6a9ukv4zhagH_6hnzvbqbHV.ceYdTcVj3IZA" rel="noopener noreferrer" target="_blank" class="">are not necessarily more likel</a><a href="https://papers.ssrn.com/sol3/papers.cfm?abstract_id=6396678&amp;__cf_chl_f_tk=RbR53Sac.RkyGfHgMS4ESEZRV5vSH7JY0jew7gashlo-1783370586-1.0.1.1-igPTvlP6a9ukv4zhagH_6hnzvbqbHV.ceYdTcVj3IZA" rel="noopener noreferrer" target="_blank" class="">y</a> than <a href="https://www.aspeninstitute.org/wp-content/uploads/2025/12/2025_Fraud-and-Scams-Analysis_Report_FINAL.pdf" rel="noopener noreferrer" target="_blank" class="">other age groups</a> to fall for financial fraud and scams. However, <a href="https://crr.bc.edu/wp-content/uploads/2014/11/wp_2014-13.pdf" rel="noopener noreferrer" target="_blank" class="">declining cognitive capabilities</a>, <a href="https://www.sciencedirect.com/science/article/pii/S0165176522000209" rel="noopener noreferrer" target="_blank" class="">overconfidence</a> in our financial capabilities, and <a href="https://www.oecd.org/en/publications/protecting-consumers-from-financial-scams-and-frauds_d41817bb-en/full-report/component-5.html#chapter-d1e479-f641103ea8" rel="noopener noreferrer" target="_blank" class="">lower levels of financial and digital literacy</a> are associated with an increased susceptibility to financial scams. Interestingly, some research suggests that people's <a href="https://psycnet.apa.org/record/2018-14865-007" rel="noopener noreferrer" target="_blank">propensity to be emotionally aroused</a> (positively or negatively) can increase the likelihood that they fall for financial scams, when making decisions in a moment of high emotion.

<h2 class="">Taking control, exiting the tunnel</h2>One <a href="https://www.researchgate.net/profile/Jurjen-Jansen/publication/326032259_Do_you_bend_or_break_Preventing_online_banking_fraud_victimization_through_online_resilience_PhD_thesis/links/5c3b027aa6fdccd6b5a8ae12/Do-you-bend-or-break-Preventing-online-banking-fraud-victimization-through-online-resilience-PhD-thesis.pdf" rel="noopener noreferrer" target="_blank" class="">interesting finding</a> from reviewing the experiences of fraud victims shows that people who fall for scams, often describe having a “gut feeling” that something is wrong, despite not being able to stop themselves from falling into the trap. Could this be the key to stopping scams before they are complete? While robust research on effective solutions is still emergent, there are three core behaviorally informed strategies that organizations and stakeholders can develop further, to help people act safely and effectively.<h3 class="">Taking a pause</h3>One way to escape an ongoing fraud or scam situation is to try to disconnect from it, even momentarily, and escape the cognitive tunnel. If the email or phone call has pushed us towards thinking automatically and with urgency, taking a pause to review what is happening can be the key to shifting to a more deliberative mindset, which could increase the chances that we notice the oddities, recognize the fraud and take a different action.Taking a pause in these situations is called an “incubation period”. <a href="https://link.springer.com/article/10.1007/s10603-026-09619-3" rel="noopener noreferrer" target="_blank" class="">Recent research</a> shows promising results for this strategy. Across multiple studies, participants who took a pause or did a different activity between reading fraudulent information and taking an action, were more likely to flag this information as fraudulent. In forcing a break in the panic to action continuum, they were <a href="https://www.sciencedirect.com/science/article/abs/pii/S2214804318304427?via%3Dihub" rel="noopener noreferrer" target="_blank" class="">more likely to consider the probabilities</a><a href="https://www.sciencedirect.com/science/article/abs/pii/S2214804318304427?via%3Dihub" rel="noopener noreferrer" target="_blank" class=""> and thus </a>less likely to fall for the scam.How, then, could financial providers help users take a pause in a scam situation? Adding deliberate friction when users take financial actions is one avenue. This might look like adding warning information in a text message before showing them the One Time Password typically requested by scammers or before they make a financial transfer. While slightly inconvenient, adding a few seconds to the user experience and a pointed reminder at the time of action could give people enough time to rethink the situation and avoid the fraud.<h3 class="">Taking control</h3>Taking a pause is an important first step, but it should be paired with a subsequent safe and effective action that protects the user. Since most fraud and scam situations start with a communication directed to the user, what if the user could then reach out to their financial institution and take control? A simple, reinforced behavior such as hanging up the phone and calling your financial institution could make all the difference.Such an instilled behavior must first be taught. Evidence reviewed across email phishing-training studies suggests that training (where fake phishing emails are sent to test users’ attention and practice flagging) can have <a href="https://www.sciencedirect.com/science/article/pii/S0167404823006053" rel="noopener noreferrer" target="_blank" class="">positive effects in the short-term</a> but may decay over time. Meanwhile, an experimental <a href="https://www.sciencedirect.com/science/article/pii/S0167268122001238" rel="noopener noreferrer" target="_blank" class="">study</a> on fraud detection training suggests that videos can help people learn how to detect fraud signals, and that reminders (training repetition) can help sustain the benefits over time.Building on the training evidence, financial organizations aiming to protect their users can create opportunities to practice the routine, such as when they go to see a teller, the next time they are speaking to customer support or via campaigns that reward repetition of the safe action (calling their financial institution after a prompt). Simplicity, repetition and feedback are the key elements for building a routine that, over time, resets the users’ automatic response to the desired behavior the next time they encounter fraud.

<h3 class="">Light in the tunnel</h3>Taking a pause and learning a safe routine behavior can help people escape a fraud or scam situation, but there’s one crucial piece left. When users are experiencing a cognitive tunnel, making sure they remember the routine or taking a pause can be difficult; their attention is focused on the urgency and the potential loss highlighted by the situation.We need to enter the tunnel by making safe cues and reminders more salient than the information presented by the scammer. <a href="https://dl.acm.org/doi/epdf/10.1145/3313831.3376709" rel="noopener noreferrer" target="_blank" class="">Research with air traffic controllers</a>, a role that requires exceptional levels of attention, shows that short but visible cues (a brief light flash that covers their full screen in orange) increased the likelihood that they paid attention to a warning message and reduced their reaction time. Could this be replicated in banking apps or platforms? What if at a random cadence, or when detecting signals that a user is potentially being scammed (atypical behavior, times of the day) the screen flashed and a message reminding them of the routine (“hang up and call us”) showed up?Similarly, if presented during the training process, visual cues can serve as a reminder of the action to be performed. <a href="https://journals.sagepub.com/doi/full/10.1177/0956797616643071" rel="noopener noreferrer" target="_blank" class="">Research shows</a> that these types of reminders work better when they are distinctive from what people would normally encounter in those situations (i.e., they are more salient). Financial institutions, then, could create their own character or symbol distinctive enough to help users remember the three key actions: take a pause, hang up and call their financial institution. Given the deafening tunnel vision in a moment of crisis, identifying where to strategically place these sirens will be crucial. For example, identifying areas of visual focus in the moment of a scam (on the credit card for when scammers are asking for card numbers, on the app log in or home pages, or right next to the confirmation button for financial transactions) allow the alarm to sound exactly at the moment of need.<h3 class="">Innovating with behavioral science</h3>Understanding people's experiences, decisions and behaviors in high-stake situations such as fraud or a scam, as well as what research shows works in similar situations, is the basis for impact-driven innovation. The proliferation and development of AI models will make it more and more challenging for users to detect frauds and scams, underscoring the ever-growing urgency for financial institutions, government and civil society to start building the guardrails, routines and systems to prevent and resolve fraud situations in real time.And remember, people’s attention is limited and precious. It is the primary resource that scammers target and the one that we must fight to protect and redirect towards prevention and fraud avoidance.

To even have a fighting chance in the battle against financial scams and fraud, in an increasingly digital and AI-driven landscape, we need to understand the challenge from a behavioral perspective.

Partners of the Common Cents Lab are tech companies, banks, credit unions, non-profits, and government organizations with the common goal of using behavioral science to improve the financial well-being of low-to moderate-income people in the United States.Startups that participate in our Startup Lab academic incubator program are exposed to behavioral economics, supported in the integration of insights from social science, and are actively involved in collaborative projects with Center For Advanced Hindsight members in order to answer startup-specific research questions.CCL is committed to sharing our knowledge about human thought, behavior, and decision-making with the world at large. Between our course material on Coursera, Udemy, and Daily Bits: we have a course for you, whether you are a hardcore behavioral science fan or are just looking to learn something new about the world around you.The Common Cents Lab is working with the Mastercard Center for Inclusive Growth to improve the financial resilience of digital platform workers and entrepreneurs in Latin America.

Common Cents Lab

Solutions that aim to increase the financial well-being for low- to moderate-income people living in the United States and abroad.

Lights in a cognitive tunnel