Riding the Big Wave: OpenAI Makes Another Splash in Higher Education with ChatGPT-4o 🌊

Article written by ErasmusX’ GenAI Team

On May 13th, 2024, OpenAI announced their newest model ChatGPT-4o (“o” stands for omni) in their spring update keynote.

TLDR: Expect higher speeds, enhanced human-machine interaction (voice and vision), and many (previously paid) features brought to ChatGPT’s free version — making them available to all.

What’s new, what’s free and what should you know if you’re in Higher Education? Spoiler alert: Higher Education will be affected to a way higher degree than ever before.

What’s New in ChatGPT-4o? 🆕

“O” for omni, refers to this version of ChatGPT being multimodal: it processes inputs and outputs (whether they are text-, voice-, or visual-based) all in one model. We know, this all sounds very technical, so here’s an explanation that actually makes sense — take it away ChatGPT: “Think of ChatGPT-4o like a super versatile tool that can understand and respond to not just text, but also voice and images. Instead of needing separate tools for each, it’s like having one Swiss Army knife that does it all. This is a big deal because it makes everything smoother and faster — imagine having to juggle different tools for different tasks versus having one handy tool that does everything seamlessly.”

Thanks Chat, we’ll take over from here again.

One thing that is immediately noticeable is that ChatGPT-4o is very fast (2x faster than its previous model), but also cheaper to run.

But the most impressive updates regard the model’s ability to deal with speech and vision. You can now talk to ChatGPT just as if it were a human being, without a keyboard, with a response time that is also human-like. You can interrupt it, ask it to respond in specific ways and make use of a very wide vocal range. All while ChatGPT is now also better at understanding what you say and at interpreting your emotions. Take a look for yourself:

Talking to ChatGPT is now very human-like

Additionally, they will also enable users to open up their cameras and use its live feed — in real time — to interact with ChatGPT about virtually everything it can see. From solving math problems to translating a foreign-language restaurant menu, anything is possible.

A better view of ChatGPT’s visual capabilities

Now, there are three more interesting practical use cases that OpenAI showed off that we did not want to keep away from you:

Realtime translation between humans

Tutoring a student in math problems

ChatGPT as one’s eyes for visually-impaired people

For now, these voice and visual features will be made available to ChatGPT Plus users first, and free users later. Additionally, OpenAI will launch a desktop app (first on Mac, later on Windows) so ChatGPT can be used in your everyday workflow. Now that it can “see”, it can help you with much, much more.

Aside from voice and visual, what other features are brought to all users? Let’s find out.

GPT-4 level intelligence and many more features coming to ChatGPT’s free version 🆓

Where ChatGPT’s most advanced features were previously locked behind a paywall, now they will become available to all users — very soon.

Here’s an overview of features that all users can enjoy soon:

• GPT-4 level intelligence: Use ChatGPT’s most advanced model (with a message limit that is 5x smaller than the paid version)
• Interact with the web: Ask ChatGPT to visit, read and analyze webpages in real time
• Upload files Ask ChatGPT to summarize, interpret and analyze documents
• Data analysis and chart creation Upload excel sheets and use ChatGPT for data-related challenges — Ideal for both researchers as well as students, but also non-academic staff
• Discover and use GPTs in the GPT Store: Build custom versions of ChatGPT, or use ones that were made by others
• Make use of the Memory feature: Make ChatGPT remember things across all chats
• Chat about photos: Upload images to discuss them with ChatGPT (ChatGPT can interpret visuals and reply to any kind of question you have about them)
• Voice and vision: Although at a later stage, it seems like the newest features will be made available to users of the free version as well (although probably at restricted capacity)

At the moment of writing (d.d. May 14th, 2024), these features are not yet made available. However, they will be rolled out in the upcoming weeks (OpenAI did not mention a specific date), so expect to be able to play with them soon!

Higher Education Has to Jump: How High? 🦘

ChatGPT-4o looks like an impressive step forward, but the biggest impact on Higher Education might not even come from its newest features. Based on 2023 app data that was reported by TechCrunch, only 0,1% of all ChatGPT users has a paid plan, keeping its most advanced features reserved for a small minority of users. In this landscape, Higher Education could afford to mostly respond to ChatGPT 3.5 and its more limited capabilities. With the new rollout on the horizon, many of the advanced features will be made available to everyone, and Higher Education will be impacted in ways it has not prepared for.

With mere 24 hours after its release it is too early for a list with best practices, but here are some early tips to get you started:

• Make use of the summer and prepare well: With the academic calendar gearing toward its end, and these features not yet made available, higher education has been granted some time to respond. Make use of the summer, whether you’re a policy maker, executive, or a teacher. Learn about ChatGPT-4o and its features, play with them, and try to gain an understanding of how they might your work within your institution. This new reality has one certainty: next academic year, all students will carry a very powerful virtual assistant with them at all times.
• Don’t lose sight of the opportunities: After the release of ChatGPT 3.5, most people — including teachers—approached it from a position of fear. But rather than reacting to it defensively, we should not forget about its opportunities. ChatGPT (and GenAI in general) can help reshape education and provide better learning experiences for students. It is not all bad news, so remember throughout the process to look for opportunities.
• Take a flexible approach: With all the new technologies that are already updated before you have the full picture, you have two options: break or bend. Perfection is impossible, and revolutions like these are always in motion. So a good response today might not work so well tomorrow — and that is fine. Working with continuously advancing technologies requires you to work with them, instead of against them—and it also helps retain your sanity 😜. Let’s find inspiration in the words of the late Bruce Lee: “Be water, my friend”.

Launching the ErasmusX GenAI Monthly Meetup 🎉

Because we know all of these challenges are quite overwhelming, we are launching the ErasmusX GenAI Monthly Meetup. We figured: why deal with these challenges all by yourself when you can work on them together?

Throughout our 4 years of existence, we have seen the power of collaboration and co-creation, especially in interdisciplinary contexts. That is why this meetup is open to all EUR teaching staff, students, and other EUR GenAI enthusiasts to come together for the purpose of knowledge sharing and gaining hands-on practice in the field of Generative AI.

👉 Interested in becoming part of the meetup? The meetups begin on May 23. Learn more about it and sign up here.