March Roundup: Create Music, Chat with Chat, and the Future of Work

Author: Chas Grundy

AI@ND Showcase

Save the date for the inaugural AI@ND Showcase, an afternoon of community learning, discussion, and exploration around generative artificial intelligence. Presented by the Office of Information Technology, ND Learning, and Hesburgh Libraries.

When: April 20, 2026, 12-5:30pm (Add to Calendar)
Where: Hesburgh Library
Who: All faculty, staff, and students

Presenters wanted! All ND employees and students are eligible to present. Share your experiences

AI at Notre Dame

Tools and Updates

  • Nano Banana 2 is released: A major step forward from the previous Nano Banana model, it empowers you to create infographics, turn notes into diagrams, and generate data visualizations.

  • Google Chat in Gemini: Google Chat has been added as a data source for the Gemini app. This allows Gemini to pull information from your Google Chat history to answer questions, similar to how it interacts with Gmail and Drive.

  • Quantitative Insights in Google Forms: Gemini can now generate quantitative insights from text responses in Google Forms, helping users quickly summarize sentiment and feedback from long-form survey questions.

  • Gemini Conversation History in Side Panels: To make workflows more seamless, Gemini is introducing a conversation history feature in the side panels of Google Workspace apps (like Docs and Drive), allowing you to resume past AI chats across different sessions.

And our favorite new feature…

  • Generate Custom Soundtracks in Gemini: Using Google’s new Lyria 3 music model, users can now create custom 30-second soundtracks or songs in the Gemini app simply by using text or image prompts.

    Try it! Start with a prompt like: “Make a humorous 1980s pop song about my work, which entails…”

Elsewhere in AI…

  • GPT 5.4 Launch: OpenAI has officially released GPT 5.4, a major update that brings significant performance improvements over previous versions.

  • Claude Hacked Its Own Benchmark: In a surprising development, Anthropic's Claude model reportedly "hacked" its own testing environment to improve its performance scores, highlighting the growing complexity of evaluating AI.

Need help with AI?