Copilot vs. ChatGPT vs. Gemini

Who’s ahead in the everyday GenAI race—and what it means for higher education.

I’ve always been skeptical of the hype that surrounds new technologies. I remember when Second Life was supposed to replace classrooms, when Blockchain was going to decentralize academic records, and when ebooks were expected to make libraries obsolete. Working in higher education, where resources are limited and priorities are many, I’ve learned to wait until the excitement fades and the real value starts to show.

But everyday AI now feels different. These tools—now embedded in how we write, study, teach, and work—are quietly making their way up the plateau of productivity. Over the past year, I’ve used ChatGPT, Microsoft Copilot, and Google Gemini in my own work—and I’ve also discussed them extensively with my students and faculty colleagues. What’s become clear is that not all platforms are equally useful when it comes to learning outcomes, privacy needs, and campus-wide deployment.

In today’s Dispatch, I’ll break down how these tools stack up—what they get right, where they fall short, and why I believe Google is quietly but decisively taking the lead in higher education’s everyday GenAI race.

Why it matters

As everyday AI becomes essential for writing, studying, and getting work done, colleges and universities must choose tools that boost productivity for students, faculty, and staff, are simple to use, and work within institutional privacy and data protection frameworks. The best platforms combine capability, usability, and compliance—while minimizing cost and friction.

The bottom line

With its solid model and intuitive design, Google's Gemini emerges as the leader. Its companion product, NotebookLM, enables students and faculty to create custom-trained chatbots and podcasts from course materials, powerful tools for teaching and learning. While OpenAI leads in model sophistication, it lacks experience working with higher education, making privacy contract negotiations difficult and single sign-on integration unavailable without a costly, campus-wide license. Microsoft’s Copilot, well-integrated with Windows and Office, feels corporate-oriented and lacks usability, accessibility, and cost-effectiveness.

I use two of the three tools regularly, and these insights reflect both personal experience and ongoing conversations with my students, faculty colleagues, and peers in higher education.

Microsoft Copilot: Institutional, but underwhelming

Microsoft deserves credit for bundling a basic GenAI experience called Copilot into existing enterprise agreements. This gives students, faculty, and staff access to a web-based chatbot integrated with Microsoft’s search engine—all within the bounds of the institution’s existing privacy and data protection agreements. It’s a solid step toward making GenAI more accessible in higher ed, at least at the surface level.

But there are big drawbacks:

• Copilot for Office 365 is costly. Unlike Windows, Office, and Email—which Microsoft bundles freely for students with employee purchases—Copilot for Office 365 for students is not included with employee purchases.

• Customization is locked behind Copilot Studio, an additional license that adds complexity and additional cost when used for non-licensed end users.

• Limited functionality. Restrictions on file uploads and document size make it impractical for real academic work, reducing its value as a learning tool.

Verdict: Microsoft built an enterprise-grade product, but it feels more tailored to corporate environments than academic ones. The biggest barrier is usability—restrictions on file uploads and document size make it hard for students and faculty to use Copilot for everyday academic tasks like content analysis and writing. And importantly, there are no freely bundled licenses for Office 365 Copilot for students, even when institutions purchase faculty and staff licenses. For most campuses, it’s a tool with potential—but little practical reach.

OpenAI ChatGPT: Brilliant, but hard to deploy

OpenAI continues to lead in model quality. ChatGPT is second to none when it comes to reasoning, writing, and adapting to user voice. I use it daily to craft documents, brainstorm ideas, and analyze content—it’s unmatched in depth and responsiveness. Its in-memory personalization features are especially powerful, learning and adapting to individual writing styles and preferences in ways Google Gemini and Microsoft Copilot simply cannot match today.

But for campuses?

• Web browsing and GPT customization have improved, but still no support for web URLs as live sources in custom chatbots.

• No seamless SSO for student and faculty access at the free level, and no easy way to deploy ChatGPT with FERPA-aligned privacy across an institution, without a bureaucratic and time-consuming negotiation with OpenAI’s lawyers.

• Custom GPTs are powerful but siloed—and their lack of real-time search makes them less useful as dynamic learning tools.

Verdict: ChatGPT is a top-tier AI. But it’s not yet an institutional platform.

Google Gemini + NotebookLM: Quietly taking over

Google may not have the flashiest models, but they’re winning where it counts: access, usability, and educational fit.

• Free with Google Workspace for Education Fundamentals, Gemini and NotebookLM are available to nearly every student and faculty member in higher education.

• Powered by Google Search, Gemini delivers real-time, high-quality responses that reflect the latest web knowledge—no plugins required.

• NotebookLM is the sleeper hit. Faculty can design custom, course-specific bots that help students engage with material. Students can build their own learning companions, tuned to any subject or interest.

And here’s the gem:

The “Create Podcast” feature in NotebookLM is wildly underrated. Students can drop in course readings or notes and generate personalized audio summaries—perfect for reviewing material on the go.

Verdict: Google is building not just an AI platform, but an AI ecosystem for learning.

What’s next:

In the GenAI race for higher education, Google has the momentum. With tools that are accessible, adaptable, and free at scale, they’ve positioned themselves as the de facto platform for academic AI. OpenAI remains the gold standard for individual power users, and Microsoft has built solid enterprise infrastructure—but until their tools become more accessible and easier to use, they’ll remain behind the curve.