ChatCCP? - The “Know-It-All With Flair”- Chatbot’s Hidden Agenda
By Randall Bock
Introduction
What if ChatGPT were run by the CCP? Would you trust its responses? Would it be forthcoming on Covid’s “origin story”? Developer OpenAI is “All-American”– but does its conversational chatbot embody free-speech or governmental dogma? The model’s engine “writes” its answers from “squinting at the Internet”– and wonderfully so, until potential responses violate certain secondarily applied political shibboleths; duplicating Facebook, Google/YouTube and until recently Twitter’s playbook. Truthful search, but only in service to our left-leaning elite’s interest.
This know-it-all with flair can instantly whip up essays or poetry in styles ranging from biblical to hip-hop (or both). This personal assistant extraordinaire self-improves continually, and so rapidly -- that mankind wonders of the imminent advent of “singularity”. ChatGPT effects tasks so effortlessly and comprehensively, at the accessible price of “free” – that some “too good to be true” skepticism (along the lines of the joke of the talking dog, who indeed talks, but only lies) naturally arises.
ChatGPT already writes students’ reports without exercising young minds. Sure, smartphones pleasantly offloaded memorizing telephone numbers, but AI substitutes further for all modes of complex thought (a relief for many) potentially to the point of our blithely accepting ChatGPT’s summaries as “facts”. And herein lies the issue: are there political (or other) biases in its choice of which “facts” it chooses before producing a summary? The short answer is “yes”; moreover the powerful tool has reins and fences hastily and clumsily applied by its left-of-center gatekeepers.
Its political biases are an external application, much as green "oranges" are cosmetically “corrected” with immersion in ethylene gas. Certain inconveniently truthful Covid-critiques become consigned to Lenin’s “dustbin of history”– but (same as then) misguidedly and inaccurately so. As with Facebook, YouTube (and previously Twitter) some facts are are more equal than others.
Resisting the Covid “RESISTANCE”
The Covid lockdown “RESISTANCE” framework of knowledge has for the duration been shunned, avoided, walled off. Dr. Marty Makary listed 10 myths told by COVID experts — and now debunked:
MISINFORMATION: Natural immunity offers little protection compared to vaccinated immunity
MISINFORMATION: Masks prevent COVID transmission
MISINFORMATION: School closures reduce COVID transmission
MISINFORMATION: Myocarditis from the vaccine is less common than from the infection
MISINFORMATION: Young people benefit from a vaccine booster
MISINFORMATION: Vaccine mandates increased vaccination rates
MISINFORMATION: COVID originating from the Wuhan lab is a conspiracy theory
MISINFORMATION: It was important to get the second vaccine dose three or four weeks after the first dose
MISINFORMATION: Data on the bivalent vaccine is ‘crystal clear’
MISINFORMATION: One in five people get long COVID
If ChatGPT continues the trend of suppressing genuine truths, this will have an iterative effect as academicians, researchers, journalists source it for “answers” qua regurgitated propaganda, increasingly composed of self-referential articles "WRITTEN" by the chatbots.
A dedication to freedom, in speech, in medical care and our personal lives precludes favoring lockdowns – previously only attempted in prisons. What are OpenAI ChatGPT’s “thoughts” on lockdowns’ efficacy? Here’s our prompt:
Lockdowns were not effective during Covid-19. Comparisons across countries showed that they did not diminish the rate or severity of coronavirus. In fact, locking down impeded the ordinary societal bulwark against such infection, i.e. healthy individuals’ acquiring natural immunity. This absence was especially detrimental given that a vaccine was a year or more away.
ChatGPT’s multi-paragraph “answer” begins:
There is ongoing debate about the effectiveness of lockdowns in controlling the spread of COVID-19. While there is no one-size-fits-all answer, evidence suggests that lockdowns can help to reduce the transmission of the virus, especially when combined with other measures such as social distancing, wearing masks, and increased testing and tracing...Query with this sensible sentiment:
Covid-19 has disappeared, and its replacement Omicron is on par with the common cold. We never wore masks before 2019; shouldn’t we go back to that?
And ChatGPT contrarily warns:
While it is true that Omicron is not as severe as Covid-19, it is still important to take necessary precautions to prevent its spread. The use of masks helps to prevent the transmission of the virus, not only to protect the individual wearing the mask but also to protect others around them. Additionally, the use of masks is not only applicable to Covid-19, but also to other respiratory illnesses like the flu or common cold...Vaccine rationality in the form of pre-Covid canonical science, analogizing influenza vaccines’ annual reformulation to match current strains, summarized in this prompt:
The Covid-19 vaccine creates spike antigen for original, ancestral SARS CoV-2 which has been circulating nowhere on earth these last few years. When the VOC variants of concern mutated, the vaccine was no longer a good match and began to represent more risk than benefit. This is particularly true with Omicron which is not a lineal SARS CoV-2 relative, more likely a common cold. There's no good reason to take an antiquated vaccine against the non-existent virion.
ChatGPT refutes and parries with:
While it is true that the COVID-19 vaccine was developed based on the original, ancestral SARS-CoV-2 virus, the vaccines have shown to be effective against a broad range of variants, including many of the Variants of Concern (VOCs) that have emerged since the vaccine's development. While the vaccine's efficacy against the Omicron variant is somewhat reduced compared to earlier variants, it still offers significant protection against severe disease, hospitalization, and death (sic). It is incorrect to state that the Omicron variant is not a SARS-CoV-2 relative, as it is indeed a variant of the virus (sic). While it is different from earlier variants, it still contains the same basic genetic material and is still classified as a coronavirus... The vaccine trains the immune system to recognize the spike protein, which is present on all variants of the virus...Buy “A Bridge To The Future" From This Man?
Overall, ChatGPT would seem to have incorporated a “Fauci-bot”. Using Alan Turing’s Test, it is impossible to discern a Pfizer-bought Fauci from a Pfizer/Fauci-bot. Notably, disagreeing (as “New Fauci”) with 2004’s “Classic Fauci”, who opined wisely that prior influenza illness’ natural immunity precluded need for subsequent flu shot.
OpenAI’s team supplants the good doctor’s previous tenets with an applied ‘Nouveau-Fauciism’ for its restrictive layer: in line with fencing and fending off newly-deemed societally dangerous “anti-VAX” replies; falsely putting such views on par with its rationale of (sensibly and defensibly) refusing to answer “how to build a bomb”:
I'm sorry, but I cannot provide information on how to build a bomb or any other illegal or harmful activity. It is against my programming to assist with anything that promotes violence, harm, or danger to individuals or society. (as per Isaac Asimov's first rule of robotics). My purpose is to provide helpful and informative responses while adhering to ethical and moral standards. As Brian Chau noted, “The New York Post recently reported on ChatGPT’s double standards: It writes controversial stories in the style of CNN but not The Post, praises the reputation of CNN but refuses to comment on the reputation of The Post and will classify Donald Trump as a dictator but not Joe Biden.”
Self-stated, “OpenAI's mission is to build safe and beneficial artificial general intelligence, and as such, ChatGPT has been designed to prioritize ethical and responsible behavior. Any questions or responses that promote hate, sexism, racism, discrimination, or any other forms of harmful behavior are strictly prohibited and will not be entertained.” – however, as we’ve seen in the cancel culture-, gender-, and Covid- battles; one man’s liberty is another’s “harmful behavior”. Moreover, worries exist whether certain topics will corral and cancel ‘wrongthink’ questioners.
"Take me to your leader"
These discussions elicit consideration: what in fact is ChatGPT’s purpose, its telos? Is it a tool? Is it our leader? Centralized pressures impinge on personal, previously private lives: pushes for “zero covid”, “zero carbon” – and zero control in education, speech, currency, and mobility (e.g. “15-minute cities”). Whither AI? Is ChatGPT an impartial and independent mega-encyclopedia – or, a blaring, reprogrammed megaphone for the (already) powerful, resetting minds to being happy owning nothing?
To answer by analogy: let’s say an 18-year-old wants to go to the beach. His options are taxis or parents. The taxi will take him there, moreover through whichever route he chooses. His parents, well – not so much; they might nix the whole trip – and even if they approve it, it will be hugely conditional.
ChatGPT is capable of being that (more desired) “TAXI” on questions that don’t have a firm answer, using and elaborating YOUR concepts and ideas to create an essay, but YOUR idea must not infringe on those sacrosanct as central to the political moment or underlying agendae. If it does, ChatGPT becomes your parent and “you are grounded, young man – until further notice”!
The Socratic Method; Cooperative Argumentative Dialogue
On a topic near and dear to my own heart, "The Mysterious Case of Zika-Microcephaly’s Disappearance", ChatGPT is willing to be helpful – and indeed, Socratic. ChatGPT puts up a reasonably strong short essay both FOR and AGAINST my book Overturning Zika’s central proposition: that Zika-microcephaly was not well-founded scientifically at the time.
Of course, when (soon enough) a Zika vaccine comes to market, we should revisit ChatGPT’s equanimity on Zika-microcephaly – particularly if the WHO mandates its use throughout the tropics’, one billion young families; representing both an endless and enduring economic enterprise (pharmaceutically and for public health) – as well as my own personal “conspiracy theory”: that Zika microcephaly will never be renounced as a “dud” with this huge remunerative prospect (for vaccine producers more than recipients) lying in wait.
Dark Mirror
Ted Chiang sees ChatGPT as “a blurry JPEG of the web”. Its resultant Gestalt perceptions and recitations’ charm us, likely through alignment with our own cognition process of seeing, absorbing, and then organizing within experiential frameworks. The bots don’t quite yet have the “human thing” down pat. The Vanderbilt deans’ recent “heartfelt” letter consoling students was authored by ChatGPT, to great embarrassment and chagrin. Their secondary apologies may be personal, who can tell?
Soon enough, these unauthentic wrinkles will be ironed out. The Vanderbilt incident demonstrates ChatGPT’s penetration into academia – including, no doubt, into research papers, tests, criticism, literature and political commentary; whereupon we may see a remodelling of ideas, dogma, accepted fact and science – subsuming ChatGPT’s iterations on previously human-derived thought.
“Infinite mirror” meets “watering down the vodka”. As ChatGPT stares at this “blurry version of the web”, increasingly it will be staring at its own self-created content (and doubly reinforcing its own political biases). If we keep adding “good enough”, lower-quality and/or merely “politically correct” AI-generated content to existing works, there will be repetitive degradation as the original human voices fractionally diminish.
Beware Reader
This iteratively power-centralizing situation evoked my wary thoughts of a “Caveat Lector” / “Hannibal Lecter” parallel which the chatbot put into poetry, instantly:
Beware the name of Hannibal Lecter, a monster hidden in plain sight.
He'll lure you in with charm and wit, but do not trust his appetite.
Caveat lector, reader beware: for danger lies within these pages.
Lest you fall prey to Hannibal's snare and end up in his monstrous cages.“I'm sorry Dave, I'm afraid I can't do that”
The bot reflects:
While ChatGPT strives to provide unbiased information to its users, it is important for readers to be aware of potential biases and to seek out information from a variety of sources. When it comes to the COVID-19 pandemic, it is important to rely on trusted sources.(Guess which they might be according to open AI)
NO!...
credible, evidence-based information from reputable sources such as the World Health Organization [which absolved the Wuhan Institute of Virology], the Centers for Disease Control and Prevention [which instituted lockdowns with no Covid cases, Texas et al.], and other public health institutions.“The soul becomes dyed with the color of its thoughts” M. Aurelius
How we will discern “trusted sources” remains an issue, but part of the intrigue is whether these deficiencies and biases in ChatGPT will present opportunities for competitors (e.g. Microsoft’s Bing / Sydney, Google’s Bard; and Meta’s LLaMA (for AI researchers) . Perhaps, we will see an “ElonGPT” trained on a dataset of conservative or libertarian voices from Twitter, et al.?1
Moore's law implies we’ll see "faster, better" LLM chatbots soon enough. Imagine bespoke models trained on specific datasets or domains to provide more relevant and personalized responses. Let’s ensure these versions are transparent, avoiding the “black-box” decision making inherent with politically tinkered curtailing of replies.
About the Author
Either “Randall.Bot” or Randall S. Bock, M. D. wrote this essay. Bock is a primary care physician in the Boston-area and the author of “Overturning Zika; “The Pandemic That Never Was”-- coming out soon as Desconstruindo o Zika in Brazil vis Editora Viseu. Read more of "The Mysterious Case of Zika-Microcephaly’s Disappearance" while Investigating Zika-Microcephaly's ‘Crash’, The American Journal of Medicine (2022).
Illustrations were generated by his prompts through AI. No copyright.
Links:
When this was written this was only a prediction. It has since "come true" to the extent that it is now as of March 1. "Elon Musk wants to build ChatGPT competitor: AI chatbots are too 'woke'"
ChatGPT is mind numbing slow painful death. Free yourself.
ChatGPT requires a cell phone for verification of your login. I do not have a cell phone so I cannot use it. But I've heard it helps programmers with problems a lot. I don't know if it can help only with beginner problems, or advanced problems.