Learn about AI Safety and Deepfakes

Gone are the days when phishing emails were easily spotted by poor spelling and grammar. With generative AI, cybercriminals can now create convincing messages, cloned voices, and realistic videos to impersonate people you trust. 

Deepfakes are becoming an increasingly sophisticated cybersecurity threat. Using publicly available photos or brief voice recordings, cybercriminals can impersonate professors, university administrators, or IT support staff to create a false sense of urgency and authority. 

What Are Deepfakes?
Deepfakes are AI-generated or manipulated images, audio, and videos designed to make people appear to say or do things they never did. Common techniques include: 

• Face Swapping: Replacing one person’s face with another. 
• Voice Cloning: Replicating someone’s voice using audio recordings. 
• Full-Body Manipulation: Altering a person’s movements or actions in a video. 

Although recognition systems can help detect deepfakes, technology is evolving so rapidly that even these quickly become obsolete, which makes it necessary to take a holistic approach. 

Using a Critical Thinking Toolkit to recognize potential deepfakes can protect you from AI-powered scams. 

How to Use a Critical Thinking Toolkit: 

• Verify Out-of-Band: If a department head or colleague sends a sudden, urgent request, stop and verify this request through an independent, known channel (like their official office number). 
• Look and Listen for Glitches: While AI is getting better, deepfakes still occasionally slip. Be on the lookout for unnatural lighting, robotic facial movements, minor audio lags, or requests that completely conflict with standard institutional policy. 
• Protect Your Footprint: The AI models feeding these scams feast on public data. Be mindful of what you post, tag, and broadcast publicly. 

Remember: Your critical thinking skills are your best firewall. Never assume digital media is genuine just because it looks or sounds familiar. 

Don’t forget: Our Cybersecurity Scavenger Hunt is still underway! Stay alert, read each week’s cybersecurity message carefully, and keep an eye out for hidden clues.  

How to Play:  
1. Read each Tuesday’s Cybersecurity Awareness email.  
2. Hunt for the hidden bolded clue letters.  
3. Save your letter. You’ll need all four weeks of clues.  
4. Solve the secret cybersecurity word at the end of the month.  
5. Submit your final answer at the end of the month using the provided Microsoft Form.  
 
Crack the code, and you will be entered into a drawing to win Bison Bucks!    

• First place: $50   
• Second place: $30 
• Third place: $20 

Good luck!