Using AI to make notebook paragraphs concise

VEX V5 General Discussion
1

Hello,

I was wondering if it was vex-legal to use AI to help make my notebook paragraphs more concise and better-said.

I am not using AI to write my notebook; I already have paragraphs of what we do day-to-day created by our notebooker, and I was wondering if it was ok to ask AI to make it sound clearer and fix all grammar/spelling issues?

2

No. There is not a single circumstance in which ai is legal in notebooks. Period.

3

From an educator point of view, it all depends on whether GDC has a AI Guidelines issued out regarding DEN.

If there isn’t any guidelines that is being communicated to the teams, then technically, nothing is stopping the teams from using AI.

And it is also wrong to say that you can’t use AI to help to detect and do auto-correction of spelling/typo errors, etc.
Even an usual MS Word or Google Doc will do that for you. Or even the WhatsApp chat or Grammerly will auto-detect it for you.
So I don’t agree with your stance on “not a single circumstance”.

It is about ethical use of AI that we are looking at. And this will include the users to cite and document all use of AI in the notebook, eg. which part was paraphrased by using AI, etc.

4

The use of artificial intelligence / large language model (AI/LLM) programs or tools to generate, organize, enhance, or alter Engineering Notebook content or programming code is contrary to the RECF Student-Centered Policy and Code of Conduct

I suppose it depends on your definition of enhance, but I feel that this paragraph from the GtJ EN4 makes it very clear that no AI should be used in the notebook.

5

As pointed out the Guide to Judging has a whole section on AI. As others have paraphrased “no”.

What we do not touch upon is the rationale is that in the Game Manual is that the robot represents the skill level of the team. You can not have pay people to make your robot or parts for you, you can not use other teams subsystems and components for your robot (even those parts machined by another team from prior season), you can not get custom cut lists from others, you can not have adults operate tools for you because you are unskilled or unsafe, you can not have others code your robot for you or use libraries you do not understand… by extension, this applies to AI generated designs, code, and documentation.

Remember this is about teams developing core skills for problem solving.

I believe it is well worth quoting the entirety of EN4 here:

<EN4> The use of artificial intelligence / large language model (AI/LLM) programs or tools to generate, organize, enhance, or alter Engineering Notebook content or programming code is contrary to the RECF Student-Centered Policy and Code of Conduct. Content produced by AI/LLM tools from prompts or by building on existing materials does not genuinely represent the skill level of a team that utilizes these tools. RECF programs offer opportunities to learn a variety of technical, organizational, and interpersonal skills. Not all students will have the same levels of competence in these skills, but all students will benefit from the practice and application of those skills as a part of the Engineering Design Process and creation of an Engineering Notebook. The misuse of AI/LLM tools, similar to non-student-centered adult involvement, deprives students of opportunities to gain experience at practicing core communication, organization, independent inquiry, and decision-making skills.

<EN5> consequences are real:

<EN5> If Judges become aware of academic dishonesty or other violations of the Student-Centered or Code of Conduct policies, those concerns should be escalated to the Judge Advisor. This may result in the removal of the team from Judged Awards at that event, and should be submitted to the RECF via the Code of Conduct reporting process for further investigation.

The underlying message is clear - Academic Honesty is paramount in VEX Robotics Competition. Taking shortcuts can lead to serious consequences that can ruin your season.

6

Thanks for pulling out this rule.

I agree with you - it all depends on the definition of enhance.
Hence, I do think GDC is trivialising this issue.
And I am saying this as the person that is overseeing the governance of the use of AI in the school.
My school’s AI guidelines run into more than 15 pages long.
It is impossible to simplify everything into just one para.

And I am genuinely interested to know how is GDC or the judging committee going to enforce the part on no auto-correction and auto-spell check.
Coz even IBO has problem in regulating this small little function.

PS. Relooking into the rule again.
Think the key phrase is “Notebook content or programming code”.
Which means as long as the use of AI is not enhancing or changing the content, then by right spell-checker and even grammer-checker should be fine.

But I am still not sure how do they know if the students made use of AI to generate the code?

7

Just want to note that the GDC does not legislate the rules for Judging/Awards Submissions. The Guide to Judging is written and upkept by the RECF Competition Judging Committee, which is an entirely separate entity from the Game Design Committee.

8

While this is debatable, what the OP said first was [quote=“Parallel15, post:1, topic:139144”]
AI to make it sound clearer
[/quote]
which would be enhancing the content of the notebook and an EN4 violation.

9

According to in the guide to judging, mistakes can be crossed out, but should not be erased completely. Mistakes are also ok to leave in…

10

I have a hard time letting this topic rest. If a team is using a Digital Engineering Notebook, let’s say Google Docs, where there is built-in support for spell-checking, does the CJC expect students to:

  1. Write out their content
  2. When making a typo or seeing the squiggly-lines for a spelling error leave the spelling error there or go back and “strike-thru” it and replace it with the fix?

While I certainly agree with the CJC that AI/LLMs can lead to similar non-student-centered adult involvement, I also believe that AI/LLMs can be done in a similar student-centered adult involvement that does lead the student to gain “organization, independent inquiry, and decision-making skills”.

In this regard, LLMs are (or can be) an “always available, expert-level, extremely patient” resource that can “level the playing field”. For example, one could construe the following interaction with OpenAI/Claude/Gemini as an EN4:

Student: Explain how PID works
LLM: Robots often need to move precisely…

To say nothing that a simple Google search now basically invokes Gemini, so does this now mean that Google searches are no longer “Student-Centered”?

I’ve had private emails with the CJC on this topic and suppose at this point am left in the “Agree to disagree” camp.

11

Thanks for pointing this out.

12

I haven’t done any notebooking that requires google searching yet, but I would just scroll down to actual sources or click the links that Gemini used to create the response. However, EN4 is about prohibiting using AI to edit/add content to the notebook, not using it to find information.

13

This may be an unpopular opinion but using a LLM can be a great teaching tool, use it that way. Here is the prompt

The next paragraph is for our robotics notebook. We can’t use AI to write for us, but can take suggestions. Don’t make changes, but tell us ongoing what kinds of things can be written better. Be critical on correctness of terms, grammar and spelling like a teacher would be.

We decided to build what is called a coaxial 4 bar arm. In a four bar linkage, 4 sides stay the same because their are a box. the box sides stay the same as they move up and down. In the coaxial linkage the two turning points are connected by gears with chains on them. as the amr spinns around the chain keeps the gears spinning together. We stole this idea by some kid named Milo. 10 uears ago.

And the response:
(297) Here’s a detailed teacher-style critique of your paragraph, with suggestions and corrections categorized by type:

:hammer_and_wrench: Technical Terminology & Accuracy

  • “Coaxial 4 bar arm”: Good attempt, but clarify the term. The correct full term is usually “coaxial four-bar linkage arm.” Also, it’s “four-bar”, not “4 bar”, in technical writing.
  • “4 sides stay the same because their are a box”: This is factually and technically incorrect. In a four-bar linkage, it’s not that the sides “stay the same” because they’re “a box”, but rather that the lengths of the four rigid links remain constant, and their relative positions define the motion.
  • “The box sides stay the same as they move up and down”: Four-bar linkages typically move in an arc, not simply “up and down.”
  • “Turning points”: The correct term is “pivot points” or “joints.”
  • “Gears with chains on them”: You might clarify this as “sprockets connected by a chain,” since gears and sprockets are different components.
  • “We stole this idea by some kid named Milo. 10 uears ago.”: “Stole” is not appropriate for a formal notebook. Also, this sentence is vague and lacks context. Consider “We adapted a design we saw from a student named Milo approximately 10 years ago.”

:blue_book: Grammar & Style

  • “their are a box” → should be “they are a box” or better: “they form a quadrilateral.”
  • “the box sides stay the same” → capitalization is inconsistent (should be “The” at the start).
  • “as the amr spinns” → should be “as the arm spins”.
  • “10 uears ago” → should be “10 years ago.”

:input_latin_letters: Spelling

  • “amr” → arm
  • “spinns” → spins
  • “uears” → years

:pencil: Style & Tone

  • Try to avoid overly casual language in a notebook, like “some kid named Milo” or “we stole.” Use academic or neutral phrasing instead.
  • Sentence structure can be clearer and more concise. Many sentences are run-ons or fragments.

As a teaching tool, the LLM are very good. Write your own words.

14

I think AI should be fine, especially since you’re just using it to clean up grammar and make things clearer. Honestly, almost everyone uses AI to format, search, or brainstorm these days. It’s kind of just a normal tool now.

15

Interesting. Trick is, how would those students write such a good prompt.

16

I’m going to assume you’re talking to the OP, as an answer to their question. If so, you are incorrect. There are no rules against using AI as a search engine (though it rarely gives correct answers to vex-specific questions), but it is not legal for any use at all in the notebook. EN4 makes it very clear that AI cannot be used to make things clearer, and using it to clean up grammar is borderline illegal.
If you weren’t talking to OP, and were instead stating an opinion, I disagree. If you want your notebook to be better, you should have to work and learn to make it better, not just feed it into an AI and paste the better version.

17

Let play a game and ask ChatGPT

And here is the answer:

(296) Great question — and it’s really smart of you to take the RECF policy seriously. To stay within the rules, your prompt should make it clear that you’re only asking for feedback or critique on your own original work, similar to how a teacher or mentor might review it.

You’re allowed to ask for help learning — you’re just not allowed to use AI to write or improve the content directly. That means you can use ChatGPT like an editor or tutor, but not as a co-author.

Here’s a prompt you can safely use:

:white_check_mark: Approved Prompt:

This text is from our engineering notebook, written entirely by our team. We’re not asking you to rewrite or generate content. Instead, act like a critical teacher and give us feedback on how we can improve the writing, grammar, use of technical terms, clarity, and professionalism — but don’t suggest specific rewritten sentences. Just point out mistakes or weak spots and explain why they’re wrong or confusing.

You could also follow up with:

Do not add or rewrite anything. Just give us honest feedback like a judge or teacher would.

:white_check_mark: What This Prompt Does:

  • Respects the RECF’s rule about student-centered work.
  • Treats ChatGPT as a critical reviewer, not a writer or editor.
  • Helps you build your own skills by understanding mistakes, not bypassing them.

:cross_mark: What You Should Avoid:

  • Asking ChatGPT to reword or improve sentences.
  • Asking it to organize or restructure content.
  • Using AI output directly in your notebook, even if it’s based on your own work.

Let me know if you want me to roleplay a “judge-style” review for a sample notebook section using this safe prompt!

There you go. Remember, these things are not “Intelligent” the have a specific framework they work in. If you ask “What is the best way to ask a question about xyz, what information do I need to add” it comes back with a pretty detailed prompt.

18

It might be interesting to ask whether it would be appropriate for one of the adults responsible for the team to do what the AI did?

19

Umm, students can learn how to write an AI prompt. It’s easier than any essay.

20

That is the hallmark of a great mentor, coach, teacher is to be able to be a critical reviewer giving both positive and negative feedback and help the roboteers improve. Read a section, offer suggestions and improvement, read the next section, offer suggestions and improvements, lather rinse repeat, the central point of iteration.

One of the huge fail points of the notebooks is that lots of roboteers never get feedback until the first judging and because of rules don’t get any then either. This is the HUGELY MISUNDERSTOOD STUDENT CENTERED POLICY. Having roboteers just blindly doing a notebook is akin to dumping a thousand dollars of parts and going “build a championship winning robot” . Nobody really wants to set their roboteer up for failure, but I see it dozens of times in a season where its clear nobody offered any improvements. So, so, so many notebooks go back with a blank rubric attached going “Hey you should look at this as a guide”.

Your question “would be appropriate for one of the adults responsible” is a resounding yes. Teaching is allowed and encouraged. You need to read all of the Policy, not just the title.

But I’ll be the first to admit, my teams have an advantage. I have decades of writing engineering documents, hundreds of thousands of words. Early in my career I was assigned to a technical editor. And through blood, sweat, tears and boxes and boxes of red ink pens, I learned how to write. And I can pass those skills onto roboteers. And while others may not have had a Judy, they can sub in an LLM. (Unlike Judy who drew tears of sadness around some of my very stilted prose, ChatGPT doesn’t do that, so the experience isn’t the same. ) Some adults are not writers, so having alternate paths is a great thing.

I’ve always been passionate about the notebooks, communication is the key to any roboteers future success. Anything that helps improve that skill I’m all in.

Thank you for attending my TED Talk.