A term recently introduced into public discourse describes the emerging – and threatening – capability of artificial intelligence – “recursive self-improvement.” Despite promising benefits of A.I. systems upgrading performance through speedy retrials, this digital technology produces an inverse, manifested in unintended, hazardous consequences – which might be called “recursive self-derangement” (RSD).
“Recursive self-derangement involves a system's self-induced disorder, potentially leading to unpredictable behaviors,” according to some online sources. (Unpredictable may prove to be a dangerously misleading euphemism.)
It seems evident that humanity itself is experiencing a similar psychological pathology, induced by over-reliance on deviant digital applications and ceaseless development of inadequately understood and therefore controllable A.I. systems.
Symptoms of RSD among humans first surged through growing use of social networking platforms. As we have subsequently learned, these platforms use algorithms that fuel rage and provoke impulsive reactions, typically propagated through deceptive and impressionistic imagery. Because using AI algorithms engages powerful primal human neural pathways, they cultivate habitual use of social networking and the spreading pathology of self-delusion that it induces.
Disturbingly, this RSD diagnosis is reinforced by the aggressive advancement of A.I., contrary to well-reasoned, insightful warnings about its dire risks. Moreover, and perhaps less obvious – this RSD diagnosis is further substantiated by humanity’s unrelenting pursuit of limitless exploitative industrialization, which is bringing Earth’s life-support systems to the brink of collapse.
We’ve become the victims of fabricated abstractions, habits of misjudgment catalyzed through deceptive digital technology – which is itself vulnerable to RSD. To prevent catastrophic RSD outcomes, rigorous Congressional control of AI is vital.