Continuing with National Cyber Security Awareness Month, a frankly unavoidable tool is AI. AI is becoming more integrated with computing systems and applications. Some of them are relatively recently released to the public while others have been around for a while. If you have a Windows 11 computer, you can probably see Copilot in the lower right of your task bar or you can always access Chat-GPT online. 

So, what can you do to lessen the risk of using generative AI insecurely? The professionals at the Cybersecurity & Infrastructure Security Agency (CISA) have created four simple principles to follow: 

1. Mind Your Inputs

AI systems learn from user inputs, so refrain from sharing anything you want to keep private, like your workplace’s company data or your personal details.

TIP: Avoid sharing sensitive or confidential information with AI models—if you wouldn’t post it on social media, don’t share it with AI.

2. Be Privacy Aware

Since AI models often scrape data from the web, what you share publicly online may be copied, in whole or in part, by AI tools.

TIP: Think about what you share with a wide audience—would you want an AI to have it?

3. Understand How Hackers Use AI

Cybercriminals may use AI to fool you. Public tools can mimic a person’s voice or image (this is sometimes called a “deepfake”). Criminals can make a voice call to mimic a trusted person and steal money or to harass people by posting fake or modified images and videos.

TIP: Stay updated on cybersecurity best practices. Criminals’ use of AI as a tool makes it more important than ever that everyone protect themselves by using The Core Four Behaviors: strong passwords, MFA, software updates, and reporting phishing.

4. Know that AI is a Tool

While AI can assist with tasks, it’s important to maintain your expertise and not rely solely on AI-generated content. Prompting isn’t the same as creating!

TIP: Treat AI as a helpful tool rather than a replacement for your skills.

---

UT Verse Vs. Microsoft Copilot

OIT offers two core AI services that support students, faculty, and staff, UT Verse and Microsoft Copilot. While general AI prompts are safe to use in Copilot, it is imperative to note that UT data should only be used in UT Verse. You can find out more about OIT’s commitment to AI on the OIT Artificial Intelligence service web pages located at https://oit.utk.edu/ai/.