Cybersecurity Awareness Month: AI Is Changing Fraud. Strengthen Verification. Reduce Risk.
Written by: Scott Morris
Cybersecurity Awareness Month is an opportunity to strengthen the everyday decisions that protect an organization’s money, information, and operations. Artificial intelligence is making it easier for criminals to impersonate executives, vendors, and colleagues. The practical response is straightforward: strengthen verification and reduce risk. Employees should not have to determine whether every convincing message, voice, or video is real. They need clear procedures for confirming sensitive requests before acting.
Consider a realistic business scenario: an employee receives a message that appears to come from an executive, followed by an AI-generated voice message that sounds like that executive, asking for an urgent payment or confidential document. The apparent follow-up makes the original request feel verified, even though both communications may come from the same attacker. This is an illustrative scenario, not a reported incident. In July 2026, the FBI warned that criminals were combining impersonation, AI-generated videos, and lookalike websites to reinforce a false sense of legitimacy. The risk is broader than one convincing fake: several apparently separate signals can support the same deception.
The safeguard is to verify the request outside the conversation the requester controls. Stop the exchange and contact the person through a trusted number or company directory already on file, not contact details supplied in the message. Confirm both the request and the person’s authority to make it, then complete the normal payment, information-sharing, or access approval process. A familiar voice, a video appearance, or a follow-up message must not substitute for that process. If the request cannot be independently confirmed, do not proceed; report it through the organization’s established channel.
Organizations across nearly every industry are adopting artificial intelligence to draft communications, analyze data, automate routine tasks, and move faster. Threat actors are capturing exactly the same benefits. The FBI warns that generative AI “reduces the time and effort criminals must expend to deceive their targets” and can correct the human errors that once served as warning signs of fraud. [2] The FBI Internet Crime Complaint Center’s (IC3) 2025 Internet Crime Report, the first to include a dedicated section on artificial intelligence, recorded 22,364 AI-related complaints and nearly $893 million in reported losses. [3]
The business impact extends beyond a convincing message. In our incident response, digital asset tracing, and cyber law work, organizations must address several questions at once: Was an account or system compromised? Who needs to be notified? Can the funds be frozen? Who may bear the loss? A sound response brings security, finance, and legal together rather than treating fraud as one department’s problem.
Attackers Can Adopt AI Without Organizational Safeguards
Responsible AI adoption requires more than selecting a useful tool. Organizations need to review how it handles information, what systems it can access, what its vendor agreements require, and who is accountable for its use. These safeguards take time, but they help prevent productivity gains from creating unmanaged business risk.
Criminals do not follow an organization’s procurement, privacy, or legal review processes. They can quickly test new tools and change their approach when something does not work. That flexibility allows them to develop and repeat deceptive requests faster.
AI adoption also creates risks inside the organization. An AI tool may hold sensitive information, connect to internal systems, or rely on outside vendors. Leaders should understand what information each tool can use and what access it has—not just how much time it saves.
A Well-Written Message Is Not Proof of a Legitimate Request
Employees have long been taught to look for misspellings, awkward wording, and unusual formatting in fraudulent emails, often called phishing. Those clues can still help, but AI can produce messages without them. Clear writing should never be treated as evidence that a request is safe.
AI tools can produce polished emails, invoices, and business correspondence in seconds, mimic industry terminology, and translate content without the errors that once exposed foreign-language scams. They can also sift through websites, social media, and company announcements to map leadership, vendor relationships, and recent events. Instead of one generic phishing email sent to thousands of recipients, attackers can send hundreds of tailored messages referencing a real conference, acquisition, or promotion.
For employees, the key question is not “Does this look real?” It is “Have I confirmed this request through our established process?” A request involving money, account access, or sensitive information should be verified independently, even when the sender appears familiar.
A Familiar Face or Voice Is Not Enough
AI-generated audio and video, often called deepfakes, can be used to imitate executives, employees, vendors, and business partners. A phone call or video meeting may feel more trustworthy than an email, but it still requires verification when the request involves a sensitive action.
The FBI’s July 2026 warning described AI-generated videos impersonating a senior FBI leader and directing people to a counterfeit complaint website. [1] For a business, the lesson is that an apparently credible person and a professional-looking destination can be parts of the same fraud. Moving from a message to a call, video, or website does not independently verify a request when the requester supplies that next step.
Public webinars, conference panels, podcasts, and videos can provide material for impersonation. Leaders do not need to withdraw from public communication. Organizations do need approval procedures that remain effective even when an employee believes they recognize the person making the request.
Regulators are seeing the same trend. The U.S. Treasury’s Financial Crimes Enforcement Network (FinCEN) has reported an increase, beginning in 2023, in suspicious activity reports describing the suspected use of deepfake media to circumvent identity verification and authentication methods. [4]
When AI-Enabled Fraud Succeeds, Who Bears the Loss?
A fraudulent request can cause a financial loss without an attacker breaking into a network. When an employee is deceived into sending money, recovery and responsibility for the loss can become complicated. Security procedures, banking agreements, and insurance terms all matter.
Commercial wire transfers in the United States are generally governed by Article 4A of the Uniform Commercial Code, rather than the consumer protection rules people may expect. Under New York’s version, a payment order is authorized if the customer approved it or is legally bound by someone acting on its behalf. An unauthorized order can also bind the customer in specified circumstances: the bank and customer must have agreed to a commercially reasonable security procedure, and the bank must have accepted the order in good faith and followed that procedure. [5]
The practical point is that a fraudulent payment is not automatically reimbursed. An employee may have instructed the bank to make the transfer, even though the employee was deceived. The organization’s ability to recover from its bank depends on the facts and the agreed security procedures. Insurance coverage also depends on policy language, coverage limits for this type of loss, and reporting deadlines.
From a security leadership perspective, payment verification is a business control with financial consequences. Finance, legal, and security should understand how their procedures work together before a fraudulent request arrives.
Report Suspected Fraud Even When No System Appears Compromised
An impersonation attempt should not remain solely with accounting, banking, or human resources because no malware or network intrusion is apparent. A deceptive payment or access request may still require a coordinated security response.
Employees should know how to report suspected fraud and should not have to prove that an account was compromised before doing so. Prompt reporting allows the response team to coordinate important steps, including:
- Preserving emails, call logs, meeting records, and payment documentation
- Determining whether an email account or system was compromised to set up the request
- Notifying insurers within policy timelines
- Evaluating whether personal or regulated information was exposed in a way that triggers legal or contractual notification obligations
Incident response plans should explicitly include suspected AI-enabled impersonation and payment fraud. Define who receives the report, who coordinates the response, and when finance, security, and legal must be involved. The response should assess both the fraud and any underlying account or system compromise.
In Payment Fraud, the First Hours Matter Most
If a fraudulent payment is discovered, act immediately. In our digital asset tracing work, funds may move through multiple accounts or be converted to cryptocurrency shortly after a transfer. Those movements can make funds harder to freeze and recover, so delays can reduce recovery options.
Contact the financial institution immediately to report the fraud and request a recall or reversal. The FBI also urges victims to file a detailed complaint at ic3.gov. Its Recovery Asset Team and Financial Fraud Kill Chain process work with financial institutions to help freeze fraudulently transferred funds. These steps support recovery efforts; they do not guarantee that the money will be returned.
Assign these responsibilities before an incident: who calls the bank, who reports to law enforcement, who notifies the insurer, and who preserves emails, meeting records, and payment details. A documented response is more useful than relying on people to make those decisions during a crisis.
AI-Enabled Fraud Belongs in Business Risk Planning
Regulators and law enforcement have now documented AI-enabled fraud in detail, including FinCEN’s November 2024 alert on deepfake fraud schemes, NYDFS guidance describing AI-enabled social engineering as one of the most significant threats to the financial services sector [6], and the FBI’s public warnings and annual tracking of AI-related complaints.
What this means: Leadership should treat AI-enabled fraud as a known business risk, not a hypothetical technology issue. After an incident, regulators, insurers, business partners, and courts may examine whether verification procedures, training, and response plans were reasonable given the available warnings. Organizations should be able to explain their safeguards and show how those safeguards are used.
Turn Awareness into Clear Responsibilities and Repeatable Actions
AI-enabled threats target business decisions as well as technology. A practical security program gives people clear rules for using AI, verifying requests, and reporting concerns. The following steps connect those rules to everyday work.
Set Clear Rules for AI Use
Identify which AI tools employees may use, what information they may enter, and who approves new tools. Keep confidential or regulated information out of unapproved AI services. Assign responsibility for vendor review and ongoing oversight. The NIST AI Risk Management Framework offers a structure for organizing this work.
Verify Sensitive Requests Through a Trusted Contact Method
For money transfers, payment instruction changes, password resets, and access requests, confirm the request through a separate, established contact method—for example, by calling a number already on file rather than one supplied in the message. Consider requiring a second approver above defined payment thresholds. Do not bypass the process because a request appears to come from an executive or is described as urgent or confidential. The FBI’s advice is to “Take a Beat” rather than act under pressure. [3]
Review Banking Agreements and Insurance Coverage
Have finance and legal review the security procedures in banking agreements and ask the bank which additional payment safeguards are available. Confirm how insurance addresses losses caused by deception, including any lower coverage limits for those losses and the deadlines for reporting them.
Teach Employees What to Do, Not Just What to Spot
Explain how AI-generated messages, voices, and videos can support impersonation. Emphasize the action employees should take: pause, verify through the established process, and report concerns. Following the procedure should not depend on noticing a technical flaw in a message or video.
Make Fraud Response Part of the Incident Plan
Include suspected AI-enabled fraud and impersonation in the incident response plan. Name the people responsible for contacting the bank, reporting to IC3, notifying the insurer, and preserving evidence. Make the reporting process clear to employees outside the security team.
Practice the Response Together
Run a discussion-based exercise, often called a tabletop, with finance, IT, legal, and leadership. Use a realistic scenario involving an impersonated executive or changed payment instructions. Test who verifies the request, who can stop the transaction, and how the issue reaches the response team. These exercises can reveal unclear responsibilities that a written plan alone may miss.
Five Things to Do Before October Ends
- Confirm that employees use a documented callback procedure, with trusted contact details already on file, to verify payment and access requests.
- Review the security procedures in your banking agreements and ask your bank what additional payment controls are available.
- Confirm whether insurance covers losses from fraudulent requests, the applicable coverage limits, and the reporting deadlines.
- Include suspected AI-enabled impersonation and payment fraud in the incident response plan, with named response contacts.
- Schedule a joint practice exercise for finance, IT, legal, and leadership using an executive impersonation or fraudulent payment scenario.
How The Beckage Firm Can Help
AI-enabled fraud requires a coordinated business, security, and legal response. At The Beckage Firm, we help clients establish AI governance, assess cybersecurity and vendor risk, strengthen incident response plans, and practice their response through tabletop exercises. When incidents occur, we support response efforts, including tracing diverted funds and pursuing recovery options.
This Cybersecurity Awareness Month, focus on actions people can follow: verify sensitive requests, use approved tools, report concerns promptly, and practice the response. The goal is not to make every employee a cybersecurity expert. It is to make secure decisions a consistent part of how the organization works.
References
[1]Federal Bureau of Investigation, Internet Crime Complaint Center, FBI Warns of Scammers Impersonating the IC3 (I-072026-PSA), IC3 (July 20, 2026).
[2]Federal Bureau of Investigation, Internet Crime Complaint Center, Criminals Use Generative Artificial Intelligence to Facilitate Financial Fraud (I-120324-PSA), IC3 (December 3, 2024), https://www.ic3.gov/PSA/2024/PSA241203
[3]FBI National Press Office, Cryptocurrency and AI Scams Bilk Americans of Billions, FBI (April 6, 2026), https://www.fbi.gov/news/press-releases/cryptocurrency-and-ai-scams-bilk-americans-of-billions
[4]Financial Crimes Enforcement Network, FinCEN Alert on Fraud Schemes Involving Deepfake Media Targeting Financial Institutions (FIN-2024-Alert004), U.S. Department of the Treasury (November 13, 2024), https://www.fincen.gov/sites/default/files/shared/FinCEN-Alert-DeepFakes-Alert508FINAL.pdf
[5]New York State Senate, Y. U.C.C. § 4-A-202, Authorized and Verified Payment Orders, NYSenate.gov (revised June 5, 2026), https://www.nysenate.gov/legislation/laws/UCC/4-A-202
[6]New York State Department of Financial Services, Cybersecurity Risks Arising from Artificial Intelligence and Strategies to Combat Related Risks, NYDFS (October 16, 2024), https://www.dfs.ny.gov/industry-guidance/industry-letters/il20241016-cyber-risks-ai-and-strategies-combat-related-risks
Cardinal News