Abstract

AI chatbots are disseminating more and more of the Internet’s search engine activity, transforming the face of education, serving as personalized AIs in intellectual and emotional relationships with humans, becoming “digital workers” that may outmode us at work, and more. Indeed, the larger category of generative AI may be one of the most transformative technologies of this decade, or even this century. Given this, it is imperative that we understand the epistemological challenges that arise with the everyday use of LLM chatbots. How will AI chatbots impact the way we come to know the world, and indeed, how will their use impact our very lives? In this piece, I articulate a major challenge that arises from their growing use, a problem which I call the “boiling frog problem.” According to the metaphor, if you boil a frog by putting it in scalding hot water, it will try to save itself. If you put the frog in a pot of tepid water, it will not notice it is boiling so it will not try to save itself. In both cases, the outcome is the same—the frog dies. In a similar fashion, the combination of factors I identify herein, over time, gives rise to unhealthy engagement with chatbots and ultimately, to diminished human agency.