AI fraud puts retail accounts at risk
On 16 November 2026, the International Organization of Securities Commissions (IOSCO) warned that regulators and industry respondents see AI-enabled fraud as the most significant emerging risk over the next 12 to 24 months. Retail traders face the greatest exposure, the organisation said.
In a statement on fraud against retail investor accounts held at market intermediaries, IOSCO described criminal groups using social engineering and digital channels to run scams across jurisdictions. Their methods include cloned broker sites, deepfakes, voice imitation and synthetic identities used during onboarding. Victims may be manipulated into sending fraudulent bank transfers or making crypto-asset withdrawals that cannot be reversed.
Broker workflows need stronger checks
IOSCO set out seven lessons for intermediaries, including brokers and trading platforms. The proposed operational focus covers stronger authentication that resists phishing, added checks when customers change account details or request withdrawals, and behavioural analytics. Login, onboarding and withdrawal processes are among the areas the statement identifies as targets.
The warning also lands as brokers add AI agents to their infrastructure. BioCatch said financial firms must distinguish genuine customers from authorised AI tools and malicious automated activity.
Its 2026 study of compliance, AML and fraud specialists across 25 nations found that four out of five institutions surveyed had encountered agentic AI threats. The findings were first reported by Finance Magnates.
| Measure | Finding |
|---|---|
| Risk outlook | AI-enabled fraud is the leading emerging risk over the next 12 to 24 months |
| IOSCO lessons | Seven operational lessons for market intermediaries |
| BioCatch study | 1,440 specialists across 25 nations, in 2026 |
| Institutions reporting agentic AI threats | Four out of five |
| AI agents named the sector's top security risk for the coming year | 84% |
| Respondents expecting difficulty separating genuine AI-enabled activity from malicious attacks | 72% |
What this means on the desk
For platform and risk teams, the alert puts customer authentication and the paths from profile changes to withdrawals in scope. Operators should check how phishing-resistant controls fit existing login and onboarding flows, and whether new checks can be applied to withdrawals without disrupting legitimate activity.
Behavioural analytics and AI-assisted activity also raise integration and review questions. Teams should establish how approved AI tools are identified, what signals feed monitoring, and how suspicious activity is escalated. IOSCO's statement does not specify implementation costs or technical requirements, so operators will need to assess those against their own systems and ask vendors what evidence supports their controls.
Platforms · Software · Providers · Markets