THE PHONY—AND THE SERIOUSLY PHONY

HAVING ONCE EXPLORED “Comparative Mendacity 101,” I feel qualified, sorta, to separate phony from seriously phony. This pertains to concepts as well as people. Here are tidbits about this perhaps unorthodox ethical position. 

Dante as Decider. As I noted back in Comparative Mendacity 101, “Like Dante, let’s arrange these sins of alternative fact in concentric circles of increasing evil.”

The key here is increasing evil. Some fabrications have largely benign effect; others are deadly serious. 

Durante degli Alighieri, known as Dante, c. 1265–1321, Italian poet. This and the following image from Wikipedia.

Flat-Earthers? Harmless Phonies, Mostly. It’s a benign belief unless the flat-earther is in a position requiring trustworthiness. A flat-earth science teacher, for instance, would be Seriously Phony.

Election Deniers? It Depends. If your MAGA uncle dominates a holiday meal, this counter-evidential guy is merely a Phony. In Dante terms, he’s qualifying himself for Limbo.

But if he manages to promote this belief among others, he’s a Serious Phony earning a place in Inferno’s inner circle.

George Santos is a Phony. Donald Trump is a Serious Phony. And neither belongs in a position requiring trustworthiness.

Artificial Intelligence. Nor has A.I. yet to have earned employment in positions requiring trustworthiness. Large Language Models scoop up vast amounts of data and apply algorithms to predict linkages. This word “predict” is important because they’ve also been known to “hallucinate”, to respond incorrectly.

That is, sometimes an A.I. response is Phony. But can it be Seriously Phony?

Annoying, But Only Phony. My Dell laptop’s default Bing has OpenAI’s GPT-4 which I find annoying to an extreme. As a test example and I admit something of an entrapment, I began a sentence with the word “How,” and it interrupts responding “are you.” If it’s so damned smart, how come it didn’t include the “?”?. Continuing my test it supplied other responses that may be statistically logical but, to my eye, they’re A.I. hallucinations.

I suspect this Bing variation got installed during an automatic update; I certainly wouldn’t have opted for it. I will be spending some time as I figure out how to disable the LLM or maybe just ditch Bing. But I stray from the point. Thus far, it’s annoying but no big deal, and therefore its interventions are Phony, not Seriously Phony.

Embodied A.I. The idea of “embodied A.I.” is to have it interact beyond the digital realm; this, by means of sensory perception and actuation. In Technopedia, December 14, 2023, Dr. Tehseen Zia describes “Embodied AI: How It Bridges the Gap Between Mind and Matter.”

Zia observes, “The foundation of our cognitive abilities as humans lies in the physical embodiment of our brains within our bodies. This physical connection enables us to interact directly with the external world. In stark contrast, most contemporary AI systems exist purely in a digital form, devoid of physical presence and lacking that immediate connection to the tangible reality we inhabit.”

Sensors and Actuators. “To achieve this,” Zia says, “embodied AI systems integrate an array of sensors, including cameras, microphones, and touch sensors, mirroring how humans rely on their senses to perceive the world.”

Image from Technopedia.

“Furthermore,” he continues, “these AI systems are equipped with actuators like wheels and motorized joints, empowering them to physically interact with and manipulate their environment, much like how we use our bodies to navigate and engage with the world around us.”

Agg! Thus far, much of A.I. can be conceptualized as “learn and inform.” In a sense, embodied A.I. is “learn and act.” And it’s this actuation that terrifies me: Embodied A.I. hallucinations could be Big-Time Seriously Phony. ds

© Dennis Simanaitis, SimanaitisSays.com, 2024