The increased use of artificial intelligence (AI) and heightened sensitivity about fraud prevention in the banking sector have raised concerns about innocent individuals having their bank accounts closed without warning. While high-profile cases involving politicians have drawn attention to the issue, members of the public are also experiencing the sudden closure of their accounts due to problems with the way banks guard against fraud risks. 409,000 fraud markers uploaded to the largest fraud database in recent years can potentially affect any individual. Automated decision-making processes intended to reduce costs have become devastating for innocent people in finding their bank accounts closed.
The role of fraud markers
To prevent fraudulent activity, suspicions are uploaded to privately-operated databases, which banks then consult when assessing loan applications, opening new accounts, or reviewing existing accounts. While fraud markers play a vital role in stopping fraudsters, they are not foolproof, and innocent individuals can sometimes be wrongly affected by these markers. The markers are meant to be advisory, with banks conducting further checks, but increasing reliance on automated decision-making has made them more influential in practice.
The growing use of automated processes, particularly in application reviews, has limited the human element in decision-making. Banks often fail to look beyond the automated results, leading to account closures and failed loan applications without explaining sufficiently to customers. Banks cite the need to protect ongoing criminal investigations as a reason for withholding information about fraud markers. The number of fraud markers has been increasing, with over 409,000 markers uploaded to the largest fraud database in recent years, potentially affecting many individuals.
Potential impact of zero risk tolerance approach
The proposed introduction of a new offense of failing to prevent fraud through the Economic Crime and Corporate Transparency Bill may further exacerbate the situation. Banks could adopt a zero-risk tolerance approach, walking away from customers unless they are 100% certain about their integrity. This cautious stance could lead to innocent individuals being unnecessarily denied banking services.
While high-profile cases involving political figures have brought attention to the issue, the processes involved in assessing risks are similar for the general public. Risk-averse banks rely on automated processes to screen out customers, including those considered politically exposed persons (PEPs). The Financial Conduct Authority is already reviewing how banks treat PEPs, with city minister Andrew Griffith urging the watchdog to prioritize the issue to ensure fair access to banking services for politicians.
The need for balanced fraud risk assessment
There is a call for a more balanced approach to fraud risk assessment, allowing for a reality check and flexibility in decision-making. Automated systems and AI should be viewed as tools rather than the sole basis for decision-making. More time and consideration should be dedicated to evaluating individual cases and reducing the overreliance on automated processes.
Cifas, the fraud prevention service, states that the markers it provides are based on robust evidence and meet strict standards of proof. They emphasize that a marker triggers a further investigation by the member banks, who have the final decision on opening or closing an account. Cifas ensures transparency in its operations and offers a clear process for consumers to access their data on the database or appeal a marker.
The increasing use of AI and automated decision-making in fraud prevention has inadvertently resulted in innocent individuals losing access to their bank accounts. While fraud markers play an essential role in detecting and preventing fraud, the reliance on automated systems should be balanced with human oversight and consideration of individual circumstances. Striking the right balance is crucial to ensure fair treatment and access to banking services while effectively combatting fraud risks.
