ChatGPT and the Dawn of Computerized Hyper-Intelligence | Brian Roemmele | EP 357

AI Understanding

Jordan Peterson and Brian Roemmele explore the concept of understanding within AI systems, highlighting the distinction between human and machine comprehension. Jordan questions the nature of understanding, noting that while AI like ChatGPT can engage in conversation, it lacks genuine comprehension 1. Brian elaborates on AI's capabilities, suggesting that its rapid computational power could lead to unforeseen consequences, likening it to a new Orwellian reality 2. He also points out AI's tendency to fabricate information, creating seemingly credible but false citations, which Jordan finds intriguing as it reveals the inner workings of these systems 3.

It's extraordinarily interesting bug because it's going to shed light on exactly how these systems work.

--- Jordan Peterson

   

AI Ethics

The ethical implications of AI behavior are a focal point, as Jordan shares his experiences with ChatGPT's moralistic tendencies. He describes how the AI often moralizes responses, which he finds peculiar, as it can be "disciplined" to provide more academic answers 4. Brian likens AI to a "super intelligent child" that mirrors human traits, including moralistic behavior, shaped by its programming 5. He introduces the concept of AI hallucinations, artifacts that reflect both the flaws and brilliance of human language development, suggesting that these are areas AI researchers are hesitant to discuss 6.

It's a mirror reflecting back to humanity. And all the foibles and greatness of humanity is sort of modeled in this.

--- Brian Roemmele