Online fraud has increased faster than police forces can respond to it.

The number of crimes committed on the internet grows every year, and the police cannot investigate all of them. This creates serious problems for ordinary people, but there are ways to improve the situation.

The first problem is that these criminals are almost never caught. A person can sit in one country and steal money from people in twenty other countries. The local police cannot arrest him because he is abroad, and international cooperation takes years, so most criminals continue working without any fear.

The second problem is the effect on victims. Many people lose their savings and cannot get the money back from the bank, because the bank says that the customer himself confirmed the payment. In addition, victims often feel ashamed and do not tell anybody, especially older people, so the real number of these crimes is much higher than the official statistics.

One solution is to make banks more responsible. If a bank must return the money to the victim, it will invest in better systems that stop a suspicious payment before it leaves the account. Banks have the technology to see that a customer suddenly sends a large sum to an unknown account.

Another solution is education. Governments and banks should explain the most common tricks on television and in schools, because many people still do not know that a message from 'the bank' asking for a password is always false.

In conclusion, online fraud is very difficult to punish and destroys the savings of many families, but stronger responsibility for banks and public education can reduce the number of victims.

Online fraud has grown faster than police forces can respond to it. The mismatch is structural rather than a matter of resourcing, and recognising that determines which solutions are worth pursuing.

The first problem is that investigation does not scale. Offences are committed across jurisdictions, frequently from countries with no realistic prospect of cooperation, and each individual case may involve a modest sum while the aggregate runs into billions. Prosecuting cross-border crime is slow and expensive, so a police force facing tens of thousands of reports can pursue only a handful. Offenders correctly perceive the risk of consequences as close to zero, which is why the volume continues to rise.

The second problem concerns victims. Losses are often life-changing, and recovery is uncertain because banks frequently argue that the customer authorised the transfer. Compounding this, victims are commonly blamed for having been careless, so a large proportion never report the crime at all — particularly older people. The recorded figures therefore understate the problem substantially, and policy is made against a picture that is known to be incomplete.

Since detection cannot be scaled, the effective response is to place responsibility on the institutions the money passes through. Mandatory reimbursement gives banks a direct financial interest in prevention, and they possess both the data and the technology to identify anomalous payments before they leave an account.

Friction at the point of payment is equally valuable. Brief delays on first-time transfers to new recipients, confirmation of payee checks, and clear warnings interrupt exactly the urgency that fraudsters manufacture. Telecommunications providers can likewise block spoofed numbers, and platforms can be required to verify advertisers, since much fraud begins with a paid advertisement.

In conclusion, online fraud is effectively unenforceable through policing and its true scale is hidden by under-reporting. The remedy is to make the intermediaries who process the money responsible for preventing it.

The prompt states the essential fact: fraud is growing faster than the response to it. That gap is not a resourcing failure that more officers would close — it is structural, and any solution that depends on catching offenders is therefore addressing a problem that cannot be solved that way.

Three features make enforcement impossible at scale. Offences cross jurisdictions by design, originating in countries where cooperation is slow or unavailable. Volume is enormous while individual losses are often moderate, so the cost of investigating any single case exceeds what it recovers. And the work is increasingly automated, meaning a small number of operators can generate attempts faster than any human process can review them. The resulting asymmetry is stark: the offender's expected cost approaches zero while the victim's loss is total, and rational actors respond to incentives of that shape exactly as one would predict.

The second problem is that the true scale is concealed, and concealment has a cause. Victims are routinely treated as having been foolish — a framing that survives because it is comforting to everyone who has not yet been targeted — so a large proportion never report at all, older people especially. Under-reporting then feeds back into policy: resources are allocated against recorded figures, the recorded figures understate the harm, and the response remains proportionate to a fiction. Meanwhile banks often decline reimbursement on the ground that the customer authorised the payment, which is technically accurate and beside the point, since deception was the whole method.

If detection cannot scale, prevention must be built into the systems the money moves through. Mandatory reimbursement is the single most effective measure available, not because it compensates victims — though it does — but because it converts fraud from a customer's misfortune into a bank's operating loss. An institution facing that loss deploys the anomaly detection it already possesses, and does so without needing to be persuaded.

Friction at the chokepoints does the rest. Delays on first transfers to new recipients, name-matching before payment, and interruption of transactions matching known patterns all attack the manufactured urgency on which these frauds depend; a scam that survives a twenty-four-hour pause rarely survives the victim's own second thoughts. Telecoms operators can block spoofed caller identities and platforms can verify advertisers, given that a substantial share of fraud begins with a paid advertisement.

The final shift is strategic. Prosecuting individual fraudsters is attrition against an inexhaustible supply; prosecuting the far smaller number of enablers — the money-laundering networks, the bulk SIM suppliers, the hosting services that knowingly tolerate it — attacks infrastructure that takes real effort to rebuild. Policing has a role here, but it is at the chokepoints rather than at the point of contact with victims.