UK APP scam losses by category, 2024
| Scam category | Loss (2024) | Share of APP losses |
|---|---|---|
| Investment | £144.4m | 32% |
| Impersonation | £101.7m | 23% |
| Purchase | £87.1m | 19% |
| Invoice and mandate | £42.7m | 9% |
| Romance | £30.5m | 7% |
| Other | ~£44.3m | 10% |
| All APP scams | £450.7m | 100% |
Investment scams are the single biggest type of authorised push payment fraud in the UK. They cost victims £144.4 million in 2024 (UK Finance), almost a third of all APP scam losses. The real story is who is exposed and how: fake platforms and AI-generated celebrity endorsements target people moving large sums, and many of these scams fall outside the new reimbursement rules.
The 2024 loss figure, its place among APP scam categories, and the AI trend reshaping how these scams are run.
Across the APP scam categories tracked by UK Finance, investment fraud takes the most money, ahead of impersonation, purchase and other scam types.
| Scam category | Loss (2024) | Share of APP losses |
|---|---|---|
| Investment | £144.4m | 32% |
| Impersonation | £101.7m | 23% |
| Purchase | £87.1m | 19% |
| Invoice and mandate | £42.7m | 9% |
| Romance | £30.5m | 7% |
| Other | ~£44.3m | 10% |
| All APP scams | £450.7m | 100% |
Investment fraud does the most damage not because it is the most common scam, but because the sums per victim are so large. The full APP picture, including the reimbursement regime, is on UK APP fraud statistics.
The mechanics are consistent: build credibility, show fake returns, then pressure the victim to send more. AI is making the credibility stage far cheaper and more convincing.
A typical investment scam starts with a plausible opportunity: a slick trading platform, a cryptocurrency tip, a "guaranteed" return. The victim sees fabricated gains on a dashboard, is encouraged to invest more, and is blocked or pressured for fees when they try to withdraw. Because the victim makes the transfers themselves, it is authorised push payment fraud, not card fraud.
What has changed is the credibility layer. AI-generated video and audio of well-known figures, so-called deepfakes, are increasingly used to fake celebrity or expert endorsements. One industry estimate attributes around 40% of biometric fraud to deepfakes. That figure is an estimate from cyber-crime reporting rather than a measured loss total, so we present it as a direction of travel, not a hard number.
It is not only individuals who are exposed.
Company directors and finance staff are targeted with fake corporate investment and treasury opportunities. The same pressure tactics work, and the sums moved from a business account can be even larger.
The mandatory reimbursement regime helps many APP victims, but several features of investment scams can put them outside its protection.
Investment fraud is measured in more than one way, and the figures are not interchangeable. The clean, like-for-like number on this page is the UK Finance APP loss.
The £144.4 million here is UK Finance's measure of investment scam losses through authorised push payments. Broader measures, such as investment fraud reported to Action Fraud, use different definitions and reporting routes and run higher, because they capture cases beyond bank-transfer APP scams. They are not directly comparable, so we keep to the UK Finance APP figure and label it clearly rather than blending sources.
Investment scam losses come from UK Finance's Annual Fraud Report; the deepfake estimate from industry fraud-and-cyber-crime reporting and is labelled as an estimate.
| Source | Publisher | Period covered | Type | Last checked |
|---|---|---|---|---|
| Annual Fraud Report | UK Finance | 2024 (calendar year) | Industry body | 3 Jun 2026 |
| Fraud and cyber-crime reporting (deepfake estimate) | Industry | Recent (estimate) | Industry estimate | 3 Jun 2026 |
The £144.4 million investment-scam loss and the full APP category breakdown come from UK Finance's Annual Fraud Report, the standard industry source.
The deepfake share is an industry estimate of a fraud technique. We label it as an estimate and keep it apart from the measured loss figures.
The 32% share and the "other" category are derived from the published APP losses and marked as derived.
Loss figures map to the UK Finance Annual Fraud Report. The deepfake share is labelled as an estimate, broader Action Fraud measures are noted as not comparable, and derived shares are flagged. Last full review: 3 Jun 2026.