ChatGPT Part II - Electric Boogaloo

What to expect:

5 min read

This week, we dive into the recent ChatGPT upgrade to see how it can help (and hinder) your students. Here’s what to expect:

• Tech Talk: Omni is here to stay

• Surplus Scoop: Washing machines - the solution we need

• Brainy Bits: ChatGPT does not equal productivity (sometimes)

Tech Talk

The next iteration is here

Last week, we saw the first major update to the world’s most used AI tool - ChatGPT. This update, dubbed GPT-4o (with the ‘o’ standing for Omni), trickles down some premium features to us humble free users. This will impact teachers, but also students, as most who are using the tool are not paying for premium. Let’s first tackle what the update brings.

The updated version now includes visual, video, and audio inputs and outputs; text is no longer the only way. This means that you can use a picture in a prompt and ask ChatGPT to produce an image of your liking and it will. On top of this, the bot overall runs significantly faster and has become even more natural in the way it responds. It now accurately picks up on emotional cues the user is giving.

Before you rush over for some new classroom art, it’s not all sunshine and roses. Free users will be limited in their response amounts. This has some implications for equability - something that ChatGPT’s parent company, OpenAI, has publicly tried to address. 

However, images and equability aside, this continues the push to have users choose ChatGPT over a simple Google search. This is where we teachers come in.  For example, language teachers can now create the wackiest image they can think of and have students use it as a writing prompt within seconds, as opposed to viewing countless message boards for the perfect quote. Science teachers can create immediate diagrams that current textbooks might not have, as opposed to scrolling on Google for endless pages until the correct cell diagram is found. Teachers are not the only users though.

With updates to ChatGPT comes renewed concerns about students cheating their way through assignments. For better or for worse, this update does not impact that side of the argument. Will cheating students get their responses a tad bit faster and will they sound more natural? Perhaps. But this does not mean that the bot will suddenly mimic how your student regularly talks or writes. The same protocols you have in place right now will still work. If we may suggest though, this may be the perfect opportunity for a teachable moment to better tackle AI-cheating with your students. Here us out:

• Even with the upgrade, ChatGPT is still not 100% accurate - let your students know this

• As a class, critique answers given by the bot so they can see its strengths and its weaknesses

• This should (hopefully) lead to a more appropriate use of it

The last point you need to know about this update is that the digital divide still continues with it. Yes, the now-free premium features are nice, but free users will only have limited access to them. This is not like internet access where speed differences don’t impact what a user can view. But instead, imagine that suddenly certain parts of every website are blocked by paywalls, or Google searches only give two results for free. This is really what is happening unfortunately with this update. Yes, the access to some premium features is a step in the right direction, but it also further increases the divide for universal access of arguably the world’s strongest educational tool. 

Will give you -4o a chance? Hit reply and let us know!

Surplus Scoop

Here’s our weekly roundup of interesting education stories from around the world. Click each link to learn more:

• Laundry machines - the surprising answer to fighting student absenteeism in New York City schools. 

• A large chunk of teachers really don’t want AI in their classrooms according to this article.

• Edtech is financially being destroyed thanks to ChatGPT.

Brainy Bits

Is it really a productivity hack?

The new ChatGPT update just came out, so it may be a while until we see some peer reviewed research on it. However, this week we look to a 2024 study from Thailand to see if AI actually makes students more productive. Here’s how they did it.

121 students from two different schools were chosen at random to complete two different tasks. The first dealt with writing/analysis and the second focused on mathematics. In each case, students were to use skills that they were unfamiliar with. In addition, each of these tasks were not simple problems, but multi-step assessments. The 121 students were divided at random between being allowed to use ChatGPT (version 3.5 at this point) and not, but all students had access to a standard basic internet browser. Comparisons were then drawn between final grades and the time it took them to complete each task. Each final grade had to be agreed upon by 3 separate teachers to help avoid bias and get the most accurate results. Ready for their findings?

At first glance, the students who could use AI universally received a better grade and in a shorter amount of time. In other words, their productivity did increase. But that’s not where this study gets interesting. If you dive into the students themselves, we get a different picture.

34% of students did not score higher or work faster than their usual work quality for writing related tasks. This number jumps to 42% for the math task. More than a quarter of students who were allowed to use AI still worked at an equivalent speed and level as they did without it. But it keeps going.

Students who were already academically successful (i.e. the top-performing students before the study) performed worse than their usual writing related submissions. ChatGPT slowed these top students down and negatively impacted their grades. Woah. But it keeps going.

Students who self-described themselves as technologically adept performed the best, regardless of their previous productivity levels. So what does this all actually tell us about student productivity and ChatGPT? 

Yes, it can make things more productive, but definitely not for every student. Students who are already succeeding in an area may have developed a set of skills and routines to tackle those types of problems; changing this up with a random AI tool won’t necessarily do them any good. Students who are academically struggling may benefit from having an extra tool, as long as they are guided appropriately on it. Regardless, the one thing this study does tell us is that if we teachers focus more on digital literacy and critical thinking skills when using tools like ChatGPT, then all types of students (should) perform at a higher level. So do just that! This way, it does not matter whatever features the next ChatGPT update brings, students will be better prepared to learn how to use it, and use it quick.

Have you noticed certain types of students perform better with AI than others? Hit reply and let us know your story!

Tell your friends about us!

Like what you’ve seen? Forward us to your friends and colleagues so y’all have something to talk about next week. 😎

Have an idea for some classroom tech, strategies, or research that you think is valuable for the community? Hit reply on this email and let us know.

Interested in growing with us? Want us to feature your tech tool, resource, or research? Hit reply to start a conversation.

Join us on social media for extra goodies, chats, and a bunch of extra resources!

References

This week’s issue adapts information from the following sources:

Tech Talk:

OpenAI. (2024). Hello GPT-4o. Retrieved from https://openai.com/index/hello-gpt-4o/ 

Brainy Bits:

Nakavachara, V., Potipiti, T., & Chaiwat, T. (2024).; Experimenting with generative AI: does ChatGPT really increase everyone’s productivity? arXiv preprint arXiv:2403.01770.