It seems that there are daily news stories about someone losing thousands or even millions of dollars through fraudulent deepfake schemes.
Deepfakes are video, audio and image content generated by artificial intelligence. This technology can produce false images, videos or sounds of a person, place or event that appear authentic. The malicious and unethical use of deepfakes can harm people by manipulating individuals or company employees and procedures. The resulting costs of this type of fraud can be very high, often destroying lifetime savings or company reserves.
Avoiding costly deepfakes requires combining awareness, technology, and procedures. Here’s a structured approach you can use to avoid becoming a victim:
Awareness and Training
- Educate employees and stakeholders about how convincing deepfakes can look and sound.
- Use real-world examples to demonstrate and train staff to spot red flags such as unusual pauses, off lip-syncing, odd backgrounds, or robotic tone shifts.
- Implement “trust but verify” culture — never act on urgent voice/video instructions involving money or sensitive data without a secondary confirmation.
Verification Procedures
- Multi-channel verification: Always confirm requests for wire transfers, data access, or confidential instructions through another trusted channel (phone call, secure app, or in-person confirmation).
- Code words or passphrases: Set up internal verification tokens for company executives so it is possible to distinguish legitimate communication from fake ones.
- Policy enforcement: Require dual approval for high-value transactions or sensitive information sharing.
Technical Solutions
- Deepfake detection tools: Deploy AI-driven detection software that flags suspicious audio/video (e.g., subtle facial mismatches, unnatural blinking, or voice timbre anomalies).
- Watermarking & digital signatures: Encourage use of content authenticity standards like C2PA or cryptographic signatures on legitimate media.
- Continuous monitoring: Monitor brand and leadership likenesses online to catch fraudulent impersonations early.
Legal and Governance Measures
- Update contracts & policies: Include clauses about verification of instructions and liability protections against impersonation.
- Incident response plan: Have a clear protocol if a suspected deepfake attack occurs—who investigates, who communicates publicly, and how losses are mitigated.
- Regulatory alignment: Stay informed about laws and standards around AI misuse and ensure compliance.
Practical Everyday Tips
- Be cautious with unsolicited video calls from executives or clients—especially with urgent financial requests.
- Train teams to spot anomalies: inconsistent lighting, blurred glasses reflections, mismatched accents, or impossible scenarios.
- Limit the amount of public audio/video of executives available online, making it harder for attackers to gather training material.
- Promote open communication: Discuss any suspicious activity with colleagues or supervisors—collaboration is one of the strongest defenses against fraud.
- Verify authenticity: If something feels off, end the interaction and reconnect through a trusted contact method to confirm legitimacy.
- Take a pause: Avoid reacting hastily to unusual or urgent requests; a moment of caution can prevent major losses.
- Stay informed: Regularly update your knowledge of new technologies and tools designed to detect and counter deepfakes.
Bottom line: Costly deepfake risks are avoided by combining people training, verification policies, and technical safeguards. The most effective defense is never relying on a single channel of trust for critical decisions.
If you have been the target of an attempted fraud, contact our office so we can make sure you, or your company have not been harmed.
Here is a step-by-step “Deepfake Defense Policy” template that you can download and share with staff.