AI in action · Investor protection
FINRA: Three Seconds of Audio Can Clone an Adviser's Voice
FINRA officials warned on November 4 that AI tools need as little as three seconds of audio to clone a voice and that an AI search engine is not a way to verify an adviser.
The scams have not changed much; the way they are delivered has. That was the message of two FINRA officials in an episode of the regulator's FINRA Unscripted podcast, reported by Traders Magazine on Wednesday, November 4. "The 'what' is very familiar, but the 'how' has fundamentally changed," said Christine Kieffer, who leads investor education at FINRA and its Investor Education Foundation. Pump-and-dump schemes and fictitious investments are old; what AI adds is speed and credibility.
The details are specific. "A modern AI tool only needs as little as three to 15 seconds of audio to clone and imitate a person's voice," Kieffer said, and a single clear photo can be enough for a basic face swap or a talking-head video. Fraudulent websites have improved considerably in the last year and a half, according to Mayur Patel, FINRA's director of financial crimes intelligence, and some now display the real Central Registration Depository number of a genuine broker or adviser, so a quick check of the number alone is not enough.
Patel's most practical warning was about the tool many people now use first. "An AI search engine is not a tool to verify stuff," he said: AI-generated answers and summaries can contain fabricated material, including about people and firms. Video calls should still be used, but they are no longer proof of identity on their own. Kieffer added that warning messages about how to spot a fake can no longer be relied on as prevention, because the fakes are too good.
The regulator's answer is to pool information. FINRA launched a Financial Intelligence Fusion Center in March, a secure portal where member firms share fraud and cybersecurity intelligence, including on criminals' use of generative AI, and offers firms a toolkit to take down websites that impersonate them or their representatives. Separately, FINRA reminded firms at its advertising regulation conference that its rules are technology-neutral, so AI-written communications face the same standards as any other, and that human review is still needed because large language models are not yet consistently accurate.
For investors the checklist is short and unglamorous. Treat urgency, guaranteed returns, exclusivity and requests for secrecy as red flags; verify a broker through FINRA's BrokerCheck by contacting the firm at a number you find yourself, not one supplied in the pitch; and talk to someone else before moving money. Patel expects losses from AI-enabled fraud to keep rising as criminals adapt, which makes slowing down the most reliable defence available.
Sources: Traders Magazine — FINRA Warns AI Is Making Investment Fraud Harder to Detect (Nov 4, 2026); FinTech Global — FINRA's AI rules expose gaps in compliance oversight (Nov 5, 2026)