285 - Retrospection

It may be time to add a new cognitive bias to the list, named “AI Retrospection”. A similar concept of “Rosy Retrospection”, the tendency to remember the past as more positive than it actually was, has been around and documented for some time. This new form of warping frequently produces laughable results, though many of them aren’t as immediately recognizable as the attached example, and some multi-billion dollar decisions have been based on equally laughable AI warping in recent memory.

As today’s over-hyped AI severely warps all content, any perceptions, conscious or unconscious, built upon that warped content are sure to manifest as a new cognitive bias among the humans whose lives are saturated with it. As this artificially induced cognitive bias takes place and takes root, it can reliably and systematically degrade the effective intelligence of any human subject to it.

“Effective Intelligence” is the product of taking any baseline cognitive ability and applying cognitive biases to it in some context, meaning that it will vary strongly based on context, the cognitive biases that are applied, and the degree to which they are applied. The net result of this is that even the highest “IQ” individual can effectively be a complete idiot if they apply cognitive biases too strongly and/or too poorly in some specific context.

Many jokes have been made about AI-hypesters “Drinking the Kool-Aide”, but that Kool-Aide is now AI-generated, giving them an infinite supply to guzzle down, tailored in a warped way to any situation. This adversarial dynamic can make the generated delusions of AI-hype as they currently circulate flow down an infinitely deep rabbit hole, which is something new and novel in the dynamics of cognitive bias.

Worse yet, this cognitive bias can also predictably impact more under-served communities more severely, such as communities whose only accessible options are often AI-generated materials, causing what is legally known as “Disparate Impact”. Ironically, tech companies frequently brag about how it is “serving” these communities in new ways, but in doing so they neglect to mention how it is systematically “mis-serving” and warping them to produce Disparate Impacts at the same time.

Other “new” cognitive biases also include the IKEA Effect and the Google Effect respectively, so perhaps this one should be named after one of the offending LLM companies. However these comically awful systems reinterpret history, history is sure to look back on them less than favorably in turn.