There are many types of AI. In this article, every mention of AI refers to the one that is trying to answer people’s questions. I am talking about ChatGPT, Google’s Gemini, DeepSeek, etc.
As a software developer, I rely on the internet’s knowledge for a living. I don’t know every answer to my problems, but I am somewhat sure that someone and somewhere on the internet have solved the problem. Hence, using search engine is my daily bread and butter.
The search engine will direct me to the relevant information in site’s documentation or some unknown blogger post. If I get lucky, I will find the answer of the exact same question in the chat forums. I often end up following the informant in other social media platforms because of their knowledge.
Most of website owners are familiar with the term search engine optimization (SEO). Traffic is good, the more the merrier. Traffic brings about business and survival. We spend hours and dollars just to ensure our site on the top list of the search engine. Search engine acts as the librarian of the internet and the origin websites remain as the knowledge supplier.
In the past 3 years or so, there is an AI boom. Apparently, you can just ask your questions to a seemingly all-knowing being. Not only software questions, but also literally everything. I have tried the tool and I am surprised by its ability to make coherent sentences. The tool can actually answer my questions correctly. But my concern is “How do they get this information? Who is their knowledge supplier?”. There is no citation at all, they seemingly derive the answers themselves.
While search engine gives traffic to knowledge supplier, AI repackage knowledge and claiming it their own. AI ability to paraphrase is astounding. It can obscure the origin of knowledge. Knowledge supplier can hardly justify whether AI answers originate from them. Making copyright strikes used to be just words to words comparison. Nowadays, you don’t even need to think about it. Everything posted on the internet can be stolen and resell by the AI trainer with zero friction.
Existing AI themselves acknowledged that their training database can be illegitimate aka stolen knowledge. I personally don’t use AI for this exact reason. AI undermines the value of being knowledge supplier. Being supportive to this type of AI is the same as being supportive to knowledge thieving.