Sorry, let me start with some self-reflection on this course.

In this class, I can see my progress. Initially, faced with the seemingly incomprehensible sample code, I instantly lost confidence in this course. There were so many codes I hadn't learned, and the usage was unclear.

Later on, I began to adapt slowly. In reality, the sample code provided by the professor is more like building blocks and a framework. Even though I didn't understand how the blocks were made, I could still use them to construct geometric shapes I liked and embed my ideas. It felt a bit like standing on the shoulders of giants. I was most pleasantly surprised by the assignment in week 3. Initially, for body tracking, I thought it would require a series of complex sensors. But unexpectedly, the AI trained could rapidly and accurately identify the human body structure, locating the parts we needed to find. I'm very satisfied with the work I produced in Week 3, feeling the joy of interacting with the computer using my hands.

As time went on, I became more proficient. After receiving feedback from classmates, I could feel that the professor slowed down a bit to ensure everyone understood. Weekly assignments no longer felt like a foggy maze. I became more willing to learn and accept new method. This class turned out to be much more flexible than I imagined. We received timely answers to various questions, tailored to different problems. I am deeply grateful for the professor's hard work and appreciate everyone's mutual assistance in the final project. The class has been evolving to adapt to the different needs of students, yet it has remained constant because we share the same goal – delving into artificial intelligence from home and gaining a deep understanding.

When it comes to my views on AI/ML, I have two points to share:

The first point is related to copyright awareness, especially as someone who follows many artists. I've noticed that a significant portion of original artists resist AI. Firstly, using others' artwork without the artist's consent and feeding it to AI is unacceptable. Artistic style is a crucial aspect. A mature art style is a personal characteristic developed over a long period of practice and technical refinement. However, because a personal style is strong, it becomes easier for AI to infer and learn the artist's style to generate works. Personally, I do not endorse the use of AI to plagiarize someone else's artistic style. An artist's style and visual content are to some extent influenced by the artist's experiences. Still, using AI to mimic someone else's style and create one's own work is like forcing a child to wear adult clothes – it's not suitable and hinders the child's move (the depth of the work).

I’m afraid this might be too absolute, but I do not support the act of feeding AI with someone else's work without their permission.

Several tools people have used to protect their art work.

Glaze https://glaze.cs.uchicago.edu/

Left side-original art work In the middle-The works generated by AI through imitation Right-the image generated by AI with the original work processing using Glaze

Nightshade, to protect work by influence AI’s understand to some specific words https://www.technologyreview.com/2023/10/23/1082189/data-poisoning-artists-fight-generative-ai/

The second one is that I first come to know about the bias in the training data. Some commonly existing discrimination and biases are likely to spread even further through AI, further solidifying people's thoughts and marginalizing certain groups. The irony is that certain beliefs held by specific groups are recognized by AI trained by them, becoming the so-called 'correct' and revered as a standard by people. So in my animation, AI is just a tool, what really matters is the human’s role playing in the AI/ML development.

reference：

Glaze website https://glaze.cs.uchicago.edu/