AI Art provided by Code.org

CS Discoveries
1

I’m not sure if I somehow missed it in previous years or if this is a new feature, but as I have students completing CSD Unit 2 Lesson 10 Project: Style a Webpage for a User, I discover that the provided images are AI generated.

Several of my students are uncomfortable with this, and rightly so. AI generated art, if it comes from Open AI’s Dall E system, has been proven to have been trained on stolen artwork. Artists have found their own watermark embedded into generated images, and some have identified other unique trademarks included. These artists were not commissioned, credited, or paid. Especially following a lesson on attributing credit only on images where the creator granted permission, I disagree with Code.org’s decision to include these images in curriculum for our students.

On top of that, the images look terrible. Take one look at any of Alana’s images and you will see a mess of letters. Savannah’s pets have some disturbing anatomy, and Kiri’s store has impossible architecture. I don’t understand how these images got past initial review, especially when Dall E has the ability to circle spots you want to fix.

Code.org is very influential for our impressionable students and should be setting the example for ethical practices online. What example are they setting if they’re taking the ethically grey “easy way out” and not taking the time to review these images before using them in a professional product? I understand they are a non-profit, and I greatly appreciate their mission to keep this education free. However, I am not convinced that they couldn’t find it in their budget to pay an artist for these images, pay for stock images, or use images already in the public domain. I request these images be replaced immediately in the spirit of supporting human-created artistic work.

2

I do not speak for code.org at all, but this ethical issue is a current “hot topic” and there are arguments on both sides of the issue. Personally, I choose to bring up the issue and make my students aware of it and let them draw their own conclusions. My wife is an artist and I am a techie who thinks that AI has a valuable role in our society, however, there certainly are limits and somewhere there’s a line that shouldn’t be crossed.

I would encourage you to send an email to support@code.org with these comments as the decision makers aren’t frequently on the forum and that kind of email can be forwarded to those who may make decisions on future implementation of AI in the courses.

Mike

3

Hi Mike,

Thanks for your reply and thoughts on the matter. I agree with you, and have done the same. I talked with my students about the ethical grey area of AI generated art specifically and encouraged them to form their own opinions and use the provided images if they so choose. I don’t have a problem with AI in general; I believe there are so many ways ML can do good for our world, and having Code.org’s internal AI grading tools is a wonderful feature. I personally draw the line at replacing human creative work with unethically trained models, especially as an artist myself. I have contacted Code.org’s support line as well since my primary concern for them is how bad they look in a professional product for our kids. I posted on here to see if anyone had the same concerns as me, or if I’m the only one bothered by it. Thanks again for reading and responding to my post.

Emma

1 Like