To outsmart modern fraud, we must first know the enemy

Fraud risk does not begin only at the point of payment, and it does not end after onboarding.

In 2025 alone, global losses related to financial fraud were estimated at $573b (US$442b), with INTERPOL warning that the threat is likely to scale significantly over the next three to five years as artificial intelligence (AI) becomes more accessible and barriers to entry remain low.

Singapore’s own experience reflects the seriousness of the threat. In a 2025 police release on the Prince Holding Group, authorities described the case as a “complex, large-scale transnational fraud network” exploiting digital and financial infrastructures across multiple jurisdictions, with police seizing or freezing more than $150m in assets.

Taken together, these figures point to something larger than rising fraud losses alone. They reflect the growth of an underground fraud economy that AI is making more structured, repeatable and easier to scale. Organisations cannot defend effectively against a threat they still fundamentally misunderstand.

Fraud rings have evolved into an industry
Fraud is still often imagined as the work of small groups or isolated bad actors. In reality, it increasingly functions like an industry, with specialist roles, service providers and repeatable operating models.

Just as a legitimate business might rely on external vendors, training partners, or specialist teams rather than doing everything in-house, fraud actors can now access tools, infrastructure, training and expertise on demand. One group may specialise in harvesting credentials, another in creating fake identities, another in moving money and another in coaching others through social engineering tactics.

No single actor needs to control the whole chain for an operation to succeed. That distributed model is precisely what makes the ecosystem so difficult to uproot.

Data sits at the centre of this economy. Much like customer data powers industries such as targeted advertising, fraud operations are increasingly built around using information to make attacks more targeted, more convincing and more effective. The more a fraudster knows about a person or business, the easier it becomes to sound believable, tailor the approach and increase the odds of success.

AI is already making organised fraud faster and more scalable
AI is not only improving how legitimate businesses work, it is also helping fraudsters save time, automate repetitive tasks, and operate more efficiently. INTERPOL has warned that AI-enhanced fraud is 4.5 times more profitable than traditional methods, showing how quickly AI is changing the economics of organised fraud.

The same technology people now use to draft emails, summarise documents, or create images can also be used to generate convincing fake identities, personalise outreach at scale and test different versions of an attack with far less manual effort.

What once required a fraudster to manually gather compromised information and stitch it together into a believable profile can now be done in a fraction of the time, producing results that are often more realistic. The result is not just more fraud, but fraud that adapts faster, scales further, and is far harder for the recipient to tell apart from something legitimate.

Organised fraud is difficult to disrupt because it is built to persist
What makes organised fraud so difficult to disrupt is its structure. It is driven less by single actors or fixed operations and more by a decentralised network that can shift across platforms, jurisdictions, and service providers. That gives the ecosystem resilience and reduces dependence on any one channel or supplier.

Even when one part of an operation is shut down, fraudsters can move to private channels, find replacement suppliers, or rely on others offering similar services, allowing operations to restart quickly. Blocking one message, account, or payment does not mean the threat has been removed. More often, it means one version of the attack has been interrupted whilst the wider network regroups and looks for its next opening.

The way forward through continuous monitoring and better coordination
The way society thinks about fraud prevention needs to change. Fraud can no longer be contained with a few fixed checks or one-off interventions. It requires earlier warning signs, more consistent vigilance. and safeguards that can adapt as threats evolve.

For businesses, the shift is toward responding more quickly to changing fraud tactics, scaling defences as they grow and protecting customers without adding unnecessary friction to the user experience. Static controls are rarely enough against threats that evolve quickly, probe for weak points and return in new forms.

Fraud risk does not begin only at the point of payment, and it does not end after onboarding. It can surface in how an account behaves, how identities are reused, how devices appear across different users or how funds move through a network. The more visibility organisations have across that lifecycle, the better positioned they are to intervene before fraud becomes loss.

At the same time, industries and wider ecosystems such as finance and fintech need stronger intelligence-sharing and closer coordination, so emerging threats, fraud rings and attack patterns can be identified earlier and acted on more collectively.

Fraud groups do not operate in silos. The response cannot afford to either.

Knowing the enemy is the first step to defending against it
The problem is not the message, the account or the transaction. It is the ecosystem behind it. The organisations that respond best will be the ones that recognise fraud as an adversary built to learn, coordinate and return, and that build their defences accordingly. Knowing the enemy is now essential. It is the first condition for staying ahead of it.