Since the system appears to be dynamic to certain extent (beyond training) this will be a permanent problem. The feedback that large/public GPT systems are give

Since the system appears to be dynamic to certain extent (beyond training) this will be a permanent problem. The feedback that large/public GPT systems are given and retrained upon will cause ups and downs, depending on the mood of the society. But I believe this entanglement of people dependent on LLMs and vice versa will go pear shaped until full Idiocracy scenario is reached. As George Carlin supposedly said "never underestimate the power of stupid people in large groups".