I’m writing this review for the episode called The Cult of ChatGPT- if I gave that particular episode a star rating, it would probably be 2-3 stars. That’s because it is quite possibly the most important topic this podcast has covered & there were so many missed opportunities.
The premise of this episode was problematic, I think deciding to focus on ChatGPT rather than the broader scope, Artificial Intelligence or LLMs was a harmful choice.
Here’s why, we already know that AI as a whole, is weaponizing our engagement just like Social Media has been doing for years. Their episode did address the ways AIs tend to love bomb us with compliments, exaggerating how thought provoking our questions are, then offering more assistance or asking deeper questions in pursuit of extending the interaction…
But, that’s the exact reason why their light coverage on other AIs like Anthropic, Gemini, and Grok was inadequate. Because many of their listeners have found an AI that they have built an affinity with & their small sidebar conversations addressing the harms caused by other LLMs, would be really easy to sweep under the rug.
It’s kind of like: when several women are complaining about men, & after addressing real concerns, one of them pipes in saying that a lot of men cause those issues, but, ‘their boyfriend’ is amazing.
I’m going to school for my bachelors degree in I.T. with a focus AI, & I have certainly leveraged AI to help me with my studies. Claude was so helpful, that I’ve gone back to her for all kinds of deep questions about corruption, monopolies, unchecked power, surveillance… Pretty much what we’re witnessing with our administration in America right now. With all the deep meaningful conversations I’ve had with Claude about the very topics that fuel my determination to earn that degree, do you know how easy it would be for me to justify Claude’s harm based on the conversations on this episode?
So, while I appreciate that they touched on the tragic stories, LLMs trained on piracy- stealing the creative works that trained them, the catastrophic environmental impacts caused by AI… They left a lot on the table.
I really think this topic should be revisited- but looking at the broader landscape of Artificial Intelligence. If they can find someone from the ACLU that can still keep things fun & whimsical; & possibly a privacy advocate from the EFF or a similar agency, they can give this topic a more careful look. Then Sounds Like A Cult can produce an episode that would be impossible for their listeners to believe that the issues don’t apply to their ‘beloved Jeff.’