Artificial intelligence is making investment scams more convincing, allowing fraudsters to impersonate financial professionals, create fake identities and build sophisticated websites in minutes, according to officials at the Financial Industry Regulatory Authority (FINRA).

Speaking on a recent episode of FINRA Unscripted, Christine Kieffer, Senior Vice President of Investor Education and President of the FINRA Investor Education Foundation, said the technology has changed how fraudsters operate, even as the schemes themselves remain familiar: “The ‘what’ is very familiar, but the ‘how’ has fundamentally changed,” she said.

According to Kieffer, pump-and-dump schemes, fictitious investments and scams built around popular market trends are nothing new. What has changed is how easily fraudsters can use AI to make themselves appear legitimate.

“Call scripts are perfected, voices are changed and impersonated, whole websites are built within minutes. Legitimate people and businesses are being cloned, and this builds source credibility for the scammer,” Kieffer said. “It’s simply a shortcut to trust,” she added.

Mayur Patel, Director, Financial Crimes Intelligence, Strategic Intelligence at FINRA, said the regulator is seeing AI used to create fraudulent identification documents, deepfake images and cloned voices. The technology is also making phishing emails, fraudulent text messages and other social engineering attacks more sophisticated.

Patel said fraudsters are using these techniques to obtain personal information, steal money or gain access to brokerage accounts. Kieffer added that traditional methods of verifying someone’s identity are becoming less reliable, particularly as AI-generated video and voice impersonations become more convincing.

Kieffer said FINRA has heard from scam victims who remained suspicious of an online contact until they agreed to a video call. Seeing the person on screen convinced them the relationship was genuine. “A modern AI tool only needs as little as three to 15 seconds of audio to clone and imitate a person’s voice,” she said.

According to Kieffer, a single clear photograph can also be enough to create a basic face swap or talking-head impersonation.

Patel cautioned that video calls should not be abandoned as a verification method, but investors should no longer treat them as definitive proof of someone’s identity. He also warned against relying on AI-powered search engines to verify investment opportunities or the people promoting them. “An AI search engine is not a tool to verify stuff. It is a more enhanced version of a Google search,” he said.

Patel warned that fraudsters can create convincing online identities and supporting material that may appear in search results or AI-generated summaries, making independent verification particularly important.

Kieffer stressed that investors also need to recognize the psychological tactics behind scams, including urgency, emotional manipulation and pressure to make quick decisions.

Kieffer pointed to urgency, promises of guaranteed returns, claims of exclusivity and requests for secrecy as warning signs that remain relevant regardless of the technology involved. “One of the most important red flags that a lot of people don’t really think about as much is a request for secrecy,” she said.

Her advice is to slow down and discuss a potential investment with someone else before committing money. That creates time to reconsider the decision and gives another person an opportunity to identify suspicious elements.

Patel said AI-enabled fraud is also creating challenges for brokerage firms, particularly as fraudsters use increasingly sophisticated impersonation techniques to target firms and their customers. He said that FINRA’s intelligence teams can identify broader patterns by combining investor complaints, tips and information submitted by member firms. “We see more of the puzzle put together rather than the pieces of the puzzle,” he said.

FINRA has been expanding its intelligence-sharing efforts as fraudsters develop new techniques. In March, the regulator launched its Financial Intelligence Fusion Center (FIFC), a secure portal through which member firms can receive and share cybersecurity and fraud-related threat intelligence. The center distributes information about emerging threats, mitigation strategies and the use of generative AI by malicious actors.

FINRA has also developed a takedown toolkit to help firms address fraudulent websites impersonating their businesses or registered representatives. Kieffer said some of these websites display legitimate Central Registration Depository (CRD) numbers, making it harder for investors to distinguish them from authentic financial services providers.

Patel said the quality of these websites has improved considerably. “If you just take a look at the websites that we were identifying maybe a year, year and a half ago, they are rudimentary compared to the websites that we see now,” he said.

Looking ahead, Patel expects losses associated with AI-enabled fraud to continue rising as criminals adapt to new safeguards. His concern is that efforts to combat fraud could receive less attention than other developments in AI, leaving investors responsible for detecting increasingly sophisticated scams.

Kieffer agreed that investor education, while important, cannot carry the burden alone. “We simply cannot rely on warning messages about how to spot a fake as a reliable prevention tool,” she said.

Instead, she argued, financial institutions, regulators and technology providers need better ways to identify and block fraudulent activity before it reaches investors. “I’m most concerned that the easy answer is to tell investors how to recognize it while that’s part of it, it’s not enough,” she said.