How To Detect AI with the Create Task

With little to no skill, a student can use AI to meet the Create Task Requirements in under 15 minutes.

IMO, I don’t see the Create Task being around in 1-2 years due to the evolution of AI sites.

Can anyone articulate if and how the AP Board is determining if a student is using AI to make a Create Task.

Future Alternatives To The Create Task: Similar to the FRQs in the AP CS A course, I see students being asked to code a few functions based upon a prompt. Example: Using a list if integers as a parameter, return the median or average or a new list of numbers which are ages 21 or higher.

Thoughts on how to deter and detect students who are going to try to use AI for their Create Task?

Thanks

In my opinion tasks like this should only be accessible during class time set out like 1 - 2 weeks worth of time for it to be done i think that would be more than enough time instead of letting them do it whenever they want you could also word salad the create task since models very much do not like that preventing a direct copy paste response, Additionally a way to point out cheating could be methods they have not used all year like [].splice() to get rid of entries that they just so happened to “use” but have not shown any use of it anywhere else (for those who are unaware this function is broadly used to remove entries from a list instead of removeItem) but i know this is more work on the people grading it so the first option would be the best one. either that or the shorter prompts with solutions could be better too instead of it spanning a week or so the test could be done in like a day with supervision

Somewhat related…I just read today that the AP Board is making changes to the other CS course, AP CS A for the 2025-2026 school year. They are changing/removing topics and also making changes to both parts of the AP Exam. IMO, the changes are making the exam easier.

So I am wondering if there will be changes to the Principles Exam as well. I haven’t heard about any upcoming changes as of yet.

1 Like

@jkeays The College Board just released the new AP CSA Guidelines and Course and Exam Description (CED).

1 Like

So true. I have told students for years that I need to see you working on it in class and I will look at your revision history. I just did another post on how code.org could make it easier on us.

1 Like

Thanks for the info.

I have also used that phrase before and actually checked but it is not convenient to anayze if they have been cheating/usingAI.

I wish all teachers were given a few stats on all coding “bubbles”

  1. % of code that was copied and pasted
  2. total time student took to complete
  3. An AI generated percent, estimating, “Code.org Believes, 85% of this code was done NOT A HUMAN”

We need tools that trigger RED FLAGS.

We also need a new set of unit end assessment questions. I hope they take the current questions and push them out as a “PRACTICE Set” since all of those answers are online. I wish we had multiple versions as well to give online based upon difficulty and another 2 versions which can easily be printed. In addition, strict rules to all teachers, if you share assessments online, we take away your account.

Overall, cheating is easier than ever before for students and I wish code.org would start addressing secure assessments, more assessment choices, secure browsers, secure coding environments and red flag trigger warnings to teachers where the platform detects AI, copy paste and cheating.

Thanks for reading

1 Like