The Rise of AI Spam

The Proliferation of Spam Information Generated by LLMs

• I like to go on Zhihu in my leisure time because I think the quality of the posts on it is relatively high.
• But recently I have found some posts on Zhihu that are well-written but lack opinions and depth.
• The style of these posts is what is called “AI at a glance”.
• It is actually relatively simple to identify these posts, that is, to see if they express opinions. Generally speaking, when a person wants to publish a post, they are inspired by something and want to express a certain point of view. Moreover, some of the big names on Zhihu either don’t write, or what they write has a certain depth and novelty, which is what I admire very much and miss very much, the so-called “legacy of Zhihu”.
• But the posts generated by LLMs are often a bunch of nonsense, without opinions, without depth, without novelty, just a bunch of fluent nonsense 🤮, and often listed in one, two, three, which looks neat in format.
• In fact, it’s not just Zhihu. On other platforms, both at home and abroad, spam generated by LLMs is beginning to flood.
• The reason for this is the fast-food-style traffic-first culture, unrestrained advertising and marketing, and malicious information manipulation.
• As an individual, what I can do is to screen some reliable sources of information for myself, compare and verify them, and maintain the vigilance of independent thinking and discrimination, but to be honest, that is quite tiring 😵‍💫.
• The current AI is far from having self-awareness. It is just a tool. After being used by people, it has a great impact on our information environment and cognition. If one day AI has self-awareness and has a level of wisdom that humans cannot match, will it formulate a strategic plan to create an all-round information cocoon for everyone to influence us, manipulate us, and brainwash all mankind?