IOSCO Warns AI Is Scaling Investment Fraud Beyond…
AI Amplifies Familiar Schemes
“Digital innovations are making fraud faster and more international than ever,” said IOSCO Board Chair Jean-Paul Servais. He pointed to timely information-sharing and cooperation across market participants, jurisdictions and IOSCO’s network as important responses. The warning concerns the industrialisation of existing techniques: technology can increase how many people a fraudster reaches and how convincingly a false identity is presented, without changing the basic objective of extracting money or account access.Impersonation exploits trust in a person or organisation, including a regulator. Account takeover compromises control of an existing account. Identity theft can support fraudulent onboarding or changes to account details. A stolen-password defence may not identify a customer willingly transferring assets after coaching by an impersonator.The distinction builds on the concerns behind IOSCO’s relationship-scam campaign. In that setting, repeated interaction can establish emotional trust before an investment request arrives. The account provider may see only the final transfer, missing the conversations that produced it. That limited visibility helps explain why neither a broker nor an investor-education campaign can address the full fraud chain alone.
A Valid Instruction Can Conceal a Fraudulent Purpose
Authentication answers who is accessing an account, subject to the reliability of the controls. It does not automatically establish whether the customer understands the destination or purpose of a transaction. For example, an investor may be persuaded to move money to what appears to be a safer account, or to release funds supposedly needed to unlock a return. Those illustrative scenarios require a different response from blocking an unfamiliar login.For intermediaries, the design question is where additional verification can be useful without making every legitimate withdrawal difficult. A newly added destination, unusual liquidation or abrupt change in account behaviour may justify a contextual check, but no single signal proves fraud. Reviews need a route to human assistance and a way to distinguish a customer making an unconventional decision from one acting under manipulation.The problem is also behavioural. The FCA’s work on pressure tactics in investment promotions illustrates why generic warnings can lose force at the moment of decision. A prompt is more useful if it addresses the actual claim being made and offers a simple way to verify it independently. Repeated boilerplate can become a habitual click-through rather than an effective interruption.
Shared Intelligence Needs Reliable Evidence
IOSCO stresses timely information exchange because contemporary schemes span products and jurisdictions. One institution may identify a suspicious destination while another sees the same actor attempting to open accounts. Combining relevant observations could improve detection, but cooperation requires clear definitions, lawful sharing arrangements and confidence in the quality of the underlying evidence. A rapid alert with a weak basis can spread error as quickly as it spreads useful intelligence.The broader direction is consistent with IOSCO’s investor-protection and AI agenda, including its international alerts infrastructure. Such tools can help identify known unauthorised operators. Their limits matter too: a recently created scam may not yet appear in a warning database, and a fraudster may impersonate a legitimate regulated firm whose name passes a superficial registration check.Commercial efforts to connect information across firms, such as Spotware’s cross-broker abuse signals, address a related visibility problem, although their purpose differs from protecting investors against external scams. That distinction must remain clear. Suspected trading-rule abuse and a customer being defrauded require different evidence, responses and remedies; a shared technical vocabulary does not make them the same risk.
Recovery and Prevention Operate on Different Timelines
Enforcement can recover assets and disrupt schemes, but usually operates after harm has occurred. The FCA’s confiscation action against a social-media-promoted Ponzi scheme illustrates the importance of asset recovery. It also underlines why prevention must act earlier, while money and account control may still be retained. Once funds cross several institutions or jurisdictions, investigation and recovery become more complex.For investors, independent verification means using contact information obtained separately from the message or person making the approach. For firms, resilience means having an accessible reporting route, preserving relevant records and coordinating promptly when suspicious activity is identified. These are operational implications of the warning, not a claim that IOSCO has imposed a new universal process. Local law, account arrangements and the circumstances of each case still determine firms’ responsibilities.The most informative performance measures would examine losses prevented, missed cases, false positives and how quickly affected customers obtain help. Counting alerts alone can reward noisy systems. Likewise, counting completed educational modules cannot establish that an investor will recognise a persuasive impersonation. The challenge is to test whether controls work together at the point where deception turns into a transfer.
