Every time I think I have an idea of what generative ai models can and can't do in the classroom, the technology moves up a notch.
That happened recently when OpenAI, creators of ChatGPT, announced that GPT-4o (the “o” stands for omni), a new and greatly improved GPT model, has been released for ChatGPT users.
The new model has faster response times and can incorporate text, video, audio and images in both its inputs and outputs, and in demos it appears to be able to converse in a natural, attention-grabbing way. technology/2024/05/chatgpt-4o-lets-you-have-real-time-audio-video-conversations-with-emotional-chatbot/” target=”_blank” data-url=”https://arstechnica.com/information-technology/2024/05/chatgpt-4o-lets-you-have-real-time-audio-video-conversations-with-emotional-chatbot/” referrerpolicy=”no-referrer-when-downgrade” data-hl-processed=”none”>Emotional cues from the person you are talking to..
Currently, GPT-4o is only available to ChatGPT Plus subscribers and select users of the free version, but OpenAI says that ultimately all users will have access to the new model, although non-subscribers will be limited in how often they can access it. that they will be able to use it.
I started using GPT-4o as a ChatGPT Plus subscriber, and while its advanced voice features aren't available yet, I'm impressed by the features that are. It's not a game changer like ChatGPT was when it was first released, but it is an improved version of the technology that will likely quickly impact education and be something that teachers will encounter in the classroom.
Why is faster GPT-4o important for teaching?
GPT-4o's faster response time was one of the new features touted in OpenAI's announcement about the technology. “It can respond to audio inputs in as little as 232 milliseconds, with an average of 320 milliseconds, which is similar to the human response time in a conversation.” the ad said.
This didn't seem like a big step forward to me on paper. Speed has never been a concern for me with ChatGPT; In fact, I've generally been impressed with how quickly it can generate text. But after using GPT-4o, I realize I was wrong!
The new model is noticeably faster, and this is important because it makes querying the ai more enjoyable and truly similar to a Google search. What's most important for educators is that increased speed is particularly important when it comes to asking multiple questions.
There has been a lot of talk about creating effective ai tutors using ChatGPT and how students can engage in dialogue with the chatbot. I'm still not convinced we're there today, but GPT-4o is definitely getting us closer by enabling more realistic, real-time conversations.
Greater visual and audio capabilities of the GPT-4o
GPT-4o can accept combinations of text, audio, image and video, and can also output text, video, audio and images. This opens up great educational potential. More voice technology will make the tool more accessible to all students, especially younger ones and those with different abilities.
While the tool's more advanced voice and video functionality has yet to be released and will be available ai/chatgpt/gpt-4o-is-openais-exciting-new-model-heres-how-to-get-access” target=”_blank” data-url=”https://www.tomsguide.com/ai/chatgpt/gpt-4o-is-openais-exciting-new-model-heres-how-to-get-access” referrerpolicy=”no-referrer-when-downgrade” data-hl-processed=”none”>In the next weeks, OpenAI's demo video shows GPT-4o having a conversation that sounds surprisingly realistic. If these capabilities prove to stand up to real-world tests, this voice feature will really boost ChatGPT's potential as an online tutor. Imagine the voice compatibility of Alexa or Siri with the best response capabilities of ChatGPT! I'm not sure if GPT-4o will get us there completely, but it seems likely to get us closer.
Meanwhile, GPT-4o's imaging capabilities can now be used in the classroom. For example, if you have access to ChatGPT Plus, you can write prompts for multiple students and then ask them to write a fictional story and create a picture of the main character or a scene.
Another fun use could be for students to critique the accuracy of the images produced. For example, I asked the model to create a realistic image of a dragon flying over New York City. Almost instantly it generated an image of a cartoon-looking dragon flying over a city that was recognizable as New York but had some major problems, including two Empire State Buildings, both located in Brooklyn rather than Manhattan.
<img alt="An ai-generated image of a dragon over lower Manhattan and Brooklyn. It mistakenly includes two Empire State buildings in Brooklyn.” srcset=”https://cdn.mos.cms.futurecdn.net/bqVBHxCkZdZbX65pybYcwN-320-80.png 320w, https://cdn.mos.cms.futurecdn.net/bqVBHxCkZdZbX65pybYcwN-480-80.png 480w, https://cdn.mos.cms.futurecdn.net/bqVBHxCkZdZbX65pybYcwN-650-80.png 650w, https://cdn.mos.cms.futurecdn.net/bqVBHxCkZdZbX65pybYcwN-970-80.png 970w, https://cdn.mos.cms.futurecdn.net/bqVBHxCkZdZbX65pybYcwN-1024-80.png 1024w, https://cdn.mos.cms.futurecdn.net/bqVBHxCkZdZbX65pybYcwN-1200-80.png 1200w” sizes=”(min-width: 1000px) 970px, calc(100vw – 40px)” loading=”lazy” src=”https://cdn.mos.cms.futurecdn.net/bqVBHxCkZdZbX65pybYcwN.png” data-pin-media=”https://cdn.mos.cms.futurecdn.net/bqVBHxCkZdZbX65pybYcwN.png”/>
Similarly, my requests to create a realistic illustration of the pier in Lake Worth Beach, Florida, resulted in an ocean pier with little resemblance to the actual pier.
Creating these types of local illustrations with students and having them identify problems can be fun. You can also make these lessons content-specific. How does GPT-4o fare when it comes to depicting D-Day scenes? What happens to the human cell?
Potential problems with GPT-4o: cheating, precision and fairness
I have written about my growing concerns about the number of ai-generated articles I see as a university professor. However, I don't think GPT-4o influences that one way or the other.
Students might generate slightly better articles a few fractions of a second faster, but that doesn't really matter, and while GPT-4o is more realistic in its short answers, I haven't seen that translate into better writing overall. . When I submitted a message to one of my classes, GPT-4o produced a response very similar to previous versions and one that would have triggered my internal ai radar as an instructor.
I have bigger concerns about GPT-4o's continued accuracy issues and less confidence as it continues to blow my mind. This is why ai education with students is so vital and why critiquing current ai capabilities with students can be a good teaching exercise. However, if I am completely honest, these problems relieve me as an educator and writer because they mean that machines cannot replace me. (At least not yet!)
My most pressing concern when it comes to GPT-4o is equity and access. Although a free version of GPT-4o will be available on a limited basis, the fact that currently only students who pay for ChatGPT Plus have access, and in the future only subscribers will have sustained access, is cause for concern. This is another sign of what I and others see as the next ai digital divide. If GPT-4o or another generative ai tool ever becomes an essential teaching tool, we must ensure that all students have access to it.
