ITS’s AI Support Team looks back on 2025’s rapid AI changes and highlights what they focused on: building AI literacy (including HawkAI courses), expanding training for enterprise tools (Copilot, prompting, Copilot Agents, and ChatGPT Edu), and helping Copilot for Microsoft 365 move from pilot to a full production service. The update also previews what’s next, continued chatbot pilots plus new efforts like a multi-model pilot and the Iowa AI Gateway, while inviting campus feedback and ideas.

This article explains five common “prompt traps” that can produce confident but unreliable outputs. Like hidden assumptions, demanding certainty, missing context, asking AI to “verify” facts it can’t check, or handing AI decisions that should stay with people. It offers safer prompt alternatives and a quick pre-submit checklist to help users get more accurate, responsible results in tools such as ChatGPT Edu and Copilot.

A myth-busting guide clarifies common misunderstandings about AI use at Iowa. Such as whether there’s only one “official” tool, whether ITS can see everything you type, and whether other Teams users can view your Copilot prompts. It also distinguishes Copilot Chat vs. Microsoft 365 Copilot and reinforces that data-classification rules still apply, especially with sensitive University data.

This piece breaks down the differences between public AI tools and Iowa-supported tools (Copilot Chat, Microsoft 365 Copilot, and ChatGPT Edu), focusing on where prompts and responses may be stored and why that matters for privacy and policy compliance. It emphasizes choosing the right tool for the data, notes that some Iowa-supported services require additional licenses, and provides simple examples of what’s safer to do in public tools vs. enterprise environments.

ITS announces that faculty and staff can purchase ChatGPT Edu licenses and explains how access works (HawkID SSO, license ordering via IT leadership, and provisioning through the University workspace). It summarizes key capabilities, training options through My Learning, and responsible-use expectations, especially around data classification, Restricted/Critical consultation requirements, and the fact that it’s not HIPAA compliant for patient data.

As the year draws to a close, we want to recognize something fundamental: every member of our university played a part in keeping the institution secure. Whether it was a faculty member pausing before clicking a suspicious link, a staff member reporting a strange login prompt, a researcher checking if a data-sharing tool met compliance needs, or a student flagging a scam circulating among peers, each action strengthened our collective ability to protect our shared mission.

When we think about cybersecurity, most of us picture external hackers trying to break in. But some of the most significant incidents begin with trusted individuals inside our community, often unintentionally. As we trust and collaborate with colleagues, it is easy to overlook how everyday actions can open the door to serious consequences.