Although I’m not paying much attention to breathless reports about imminent strong AI, the Singularity, and computers already able to “model” human cognition and perform “impressive” feats of creativity (e.g., responding to prompts and creating “artworks” — scare quotes intended), recent news reports that chatbots are harassing, gaslighting, and threatening users just makes me laugh. I’ve never wandered over to that space, don’t know how to connect, and don’t plan to test drive for verification. Isn’t it obvious to users that they’re interacting with a computer? Chatbots are natural-language simulators within computers, right? Why take them seriously (other than perhaps their potential effects on children and those of diminished capacity)? I also find it unsurprising that, if a chatbot is designed to resemble error-prone human cognition/behavior, it would quickly become an asshole, go insane, or both. (Designers accidentally got that aspect right. D’oh!) That trajectory is a perfect embodiment of the race to the bottom of the brain stem (try searching that phrase) that keeps sane observers like me from indulging in caustic online interactions. Hell no, I won’t go.
The conventional demonstration that strong AI has arisen (e.g., Skynet from the Terminator movie franchise) is the Turing test, which is essentially the inability of humans to distinguish between human and computer interactions (not a machine-led extermination campaign) within limited interfaces such as text-based chat (e.g., the dreaded digital assistance that sometimes pops up on websites). Alan Turing came up with the test at the outset of computing era, so the field was arguably not yet mature enough to conceptualize a better test. I’ve always thought the test actually demonstrates the fallibility of human discernment, not the arrival of some fabled ghost in the machine. At present, chatbots may be fooling no one into believing that actual machine intelligence is present on the other side of the conversation, but it’s a fair expectation that further iterations (i.e., ChatBot 1.0, 2.0, 3.0, etc.) will improve. Readers can decide whether that improvement will be progress toward strong AI or merely better ability to fool human interlocutors.
Chatbots gone wild offer philosophical fodder for further inquiry into ebbing humanity as the drive toward trans- and post-human technology continue refining and redefining the dystopian future. What about chatbots make interacting with them hypnotic rather than frivolous — something wise thinkers immediately discard or even avoid? Why are some humans drawn to virtual experience rather than, say, staying rooted in human and animal interactions, our ancestral orientation? The marketplace already rejected (for now) the Google Glass and Facebook’s Meta resoundingly. I haven’t hit upon satisfactory answers to those questions, but my suspicion is that immersion in some vicarious fictions (e.g., novels, TV, and movies) fits well into narrative-styled cognition while other media trigger revulsion as one descends into the so-called Uncanny Valley — an unfamiliar term when I first blogged about it though it has been trending of late.
If readers want a really deep dive into this philosophical area — the dark implications of strong AI and an abiding human desire to embrace and enter false virtual reality — I recommend a lengthy 7-part Web series called “Mere Simulacrity” hosted by Sovereign Nations. The episodes I’ve seen feature James Lindsay and explore secret hermetic religions operating for millennia already alongside recognized religions. The secret cults share with tech companies two principal objectives: (1) simulation and/or falsification of reality and (2) desire to transform and/or reveal humans as gods (i.e., ability to create life). It’s pretty terrifying stuff, rather heady, and I can’t provide a reasonable summary. However, one takeaway is that by messing with both human nature and risking uncontrollable downstream effects, technologists are summoning the devil.