Here’s a new way to lose an argument online: the appeal to AI
Here’s a new way to lose an argument online: the appeal to AI
The author discusses the growing phenomenon of the 'appeal to AI,' a logical fallacy where individuals reflexively trust and cite outputs from large language models like ChatGPT in arguments. This is explored through examples ranging from seeking medical advice to philosophical debates, highlighting the issue of relying on technology that generates plausible but not necessarily trustworthy or accurate information.
Note
an understandably strong negative reaction. yet, it's a slippery slop and could be applied to people repeating verbatim answers from google search or calculators.
the articles angst definitely points to a number of issues with LLMs today (and for that matter, search results, 24 hour news talking heads, podcasters, etc etc). though, i think a number of solutions (though complicated in their own right) are possible: providing context, re-evaluating answers grounded in trust worthy resources, getting a second opinion, avoid taking an authoritative voice, etc etc.
the solutions aren't simple and i've seen some take nice strides in those directions as well.
though we may have mis-aligned incentives today that are not rewarding such solutions and instead magic is the preferred method...
Published
Summary
Image
