Educational Rumble: ChatGPT

Unless you’ve been living under a rock the last few weeks, you’ve probably heard of this new thing that's taken the world by storm- ChatGPT.

The Robots are coming...

Yet, many students seem to use it in interesting ways- ways that teachers don’t like at all. In fact some schools are actually starting to ban the technology. NYC public schools will "ban both teachers and students from using ChatGPT".

This is gonna get interesting.

At the risk of sounding like a surfer, I must say that educators should flow with the wave. Don’t fight the wave- that never works. 

Imagine if google was banned in schools since 2000? Oh wait, they did. A lot of schools in the 2000’s banned both google and the internet, but then came to realize their mistakes over time. 

Schools all over the country might start banning this AI technology, but this may only increase the reward for those who know how to cover up the fact that they used the technology. You can’t stop hackers.

These language models are quite literally trained from all the text on the internet and output text based on input prompts. Moreover, the copyright issues around the machine-generated text are still in a gray area.

I propose that teachers think about this technology in terms of the inputs in a system in comparison to the outputs- let me explain. 

Though the models are trained from the corpus of text on the internet, the prompt that a given student inputs into the system do require some level of thought- so, they should probably be graded, in addition, on the quality of this input.

As Elon Musk once stated, “One of the really tough things is figuring out what questions to ask. Once you figure out the question, then the answer is relatively easy”. We can apply Elon’s insight to the AI problem in schools. How good is a given student’s input into the system?

Anyway, it's also important to consider that ChatGPT is not perfect by any means. For Computer science students who code in SQL, professors at Stanford have found that AI-written SQL injections produced “more security vulnerabilities”. These Stanford professors also found that students who used the AI assistant in writing code felt more confident about it, even if it was more vulnerable to attacks. 

So, teachers can sit back and breathe- the AI your students use is NOT perfect, so a student’s brain must still be utilized!

Hear that, snitches?

Anyways, I want everyone to go to ChatGPT write now and input “give me the perfect vanilla latte recipe”. Don’t feel like it? 

Well here’s the output below:

Here is a recipe for a classic vanilla latte:

Ingredients:

• 1 cup of milk

• 1 shot of espresso

• 1 tsp vanilla extract

• 1-2 tbsp sugar (optional)

Instructions:

1. Heat the milk in a saucepan over medium heat until it begins to steam. Alternatively, you can use a milk frother to froth the milk.

2. Pull a shot of espresso into a mug.

3. Add the vanilla extract and sugar (if using) to the mug. Stir to combine.

4. Pour the frothed milk over the espresso mixture.

5. Serve immediately and enjoy!

You can also top the latte with a sprinkle of cinnamon or cocoa powder for an added touch of flavor.

Saved you like $5, if you were gonna go to Starbucks.

Here's a joke by ChatGPT: 

"Why do programmers always mix up Halloween and Christmas? Because Oct 31 equals Dec 25."

Funny, right? Right?

Anyways, peace out for today, lads.