Cybersecurity Awareness Fact Sheet
For public distribution – Updated June 2024
Generative AI tools like ChatGPT support new content creation and streamlined workflows, but they also pose cyber risks. Cybercriminals use AI tools to conduct sophisticated phishing and impersonation schemes.
Spotting AI-Generated Characteristics
- AI-generated phishing messages may sound more credible, personalized, and lack spelling and grammar errors.
- Look for contextual errors, misinterpretations, and unusual phrasing within messages.
- Pay attention to tone of voice and content. AI-generated text may lack emotion and contain repetitive words or ideas.
- While photos and videos might look normal at first, AI deepfakes often contain image distortion and inconsistencies (e.g., body parts out of place, strange objects), as well as textured or airbrushed backgrounds.
- In AI-generated videos, shadows or lights may flicker and different aspects of the video may blend together and shift irregularly.
- To identify AI-generated audio, listen closely for atypical background noise and unnatural speech and pronunciation.
Best Practices
- Consider the context of a video, call, or image. If the sender is pressuring you to take an urgent action, step back and ask yourself if it is a scam.
- Use caution when answering unsolicited phone calls, emails, or other online messages from individuals asking for sensitive information or featuring a call to action. Even if it appears to come from a familiar contact, a scammer could have used AI tools to impersonate them.
- If you suspect a photo or video might not be authentic, look for AI-generated characteristics and conduct a reverse image search to see if it’s been shared on other websites.
Reporting & Resources
- Report suspicious emails received on your work email to your organization’s appropriate security teams.
- Report suspicious messages received on personal accounts to the platform and the Federal Trade Commission.
Diplomatic Security Service
U.S. Department of State
