Ai, Ai, Ai Caramba!

I am certain Meta Ai is generating increasing numbers of FakeBook texts.

I'm noticing something new and a bit disturbing, and wonder if I am the only one.

On my FecesBook feed, I am getting more and more links to sites I never followed, but may have responded to someone's post there, or the site is similar to sites I AM following. Indeed, each one has a Follow button, to get me to receive more from the site regularly.

I also notice that each one has a Meta link after it with some "start a conversation" button relating to the text in the feed itself.

But here is what I am noticing in EACH of these particular feeds:

There is some long text relating to an image coming from the site. Very long text. And something in EACH text seems to be of the same style and format as ALL the other texts. That is, they all seem to be authored by the same source, and NONE OF THEM READS like a normal human writer!

I suspect each is the product of Ai, specifically Meta Ai, as sort of a come-on for the sites to USE Meta Ai to save time and give the text from the site the "benefit" of the supposed scope and "polish" that the site's admins believe they may lack.

If so, this is a big mistake on the admins' part and a dreadful sign of things to come.

As a one-time professional editor, I got good at recognizing styles and the "tells" that certain writers employed, often unwittingly. As an example, when I used to comment on posts at the site of the local weekly newspaper I was a contributing columnist to (also online, which allowed for far more input than what a letters to the editor space in the print version could ever allow), there were a few persons who, to try to escape my pummeling of them, would use pseudonyms or post anonymously. But I always saw through their attempts, because their style -- their grammar, spelling, or use of certain phrases (one person always referred to "mom and pop stores" as "mama papa stores") -- was too consistent, even without posting their real names, to stay hidden for long to a trained eye.

Even when I tried to do likewise, posting anonymously to avoid being attacked by the few who were obsessed WITH attacking me, and trying to "dumb down" my style, it was impossible, because I was still the only person who was consistently logical, erudite, and was ruthless in responding to EVERY point in the other persons' posts.

Are you aware of the Turing Test? One of the inventors credited with ushering in the computer age, Alan Turing devised a test to apply to so-called "smart" computers:

The Turing test, originally called the imitation game by Alan Turing in 1949, is a test of a machine's ability to exhibit intelligent behavior equivalent to, or indistinguishable from, that of a human. Turing proposed that a human evaluator would judge natural language conversations between a human and a machine designed to generate human-like responses. The evaluator would be aware that one of the two partners in conversation was a machine, and all participants would be separated from one another. The conversation would be limited to a text-only channel, such as a computer keyboard and screen, so the result would not depend on the machine's ability to render words as speech. If the evaluator could not reliably tell the machine from the human, the machine would be said to have passed the test. The test results would not depend on the machine's ability to give correct answers to questions, only on how closely its answers resembled those a human would give. Since the Turing test is a test of indistinguishability in performance capacity, the verbal version generalizes naturally to all of human performance capacity, verbal as well as nonverbal (robotic). (from Wikipedia)

Some claim that the Turing test has been passed and that it is now possible for Ai software/hardware to fool persons to the point that one can not tell if the "author" of the responses is human or machine.

This may be so for some examples, but I am suggesting that the examples I am seeing on FakeBook are full of too many "tells" that give away the Ai nature of their creation. I'm sure that, with time, just as movie CGI and image Ai has improved to the point where images and videos can now be created that are nearly impossible to tell from authentic naturally-created ones, such texts, which I'm sure are Meta Ai-generated, may learn how to hide these tell-tale affectations.

What do I mean? Well, first of all, these texts tend to be too long. Due to repetition and needless extrapolation, as if to preen over how much "it" knows, the texts' "authors" don't understand the value of brevity and getting to the point and being done with it. There's an old adage for public speaking that goes, "Tell them what you are going to tell them, tell them, tell them what you told them." These texts take that adage too far, and belabor the obvious points already made. This flaw is seen in nearly every example I've come across.

Second, when it makes a glaring error in facts, it forges ahead, never suggesting that it may be wrong or that it's not sure. And often these "facts" conform to a DEI woke perspective. For example, one text about making the original Matrix film, said the creators were "Lana and Lilly Wachowski." Now, when the film was made, LARRY and ANDY Wachowski hadn't yet pretended to be women, and the names THEY WERE BORN WITH appeared on the film. THAT is factual history. It is Orwellian to change the past facts to conform to later revisions.

ONLY a woke-driven program would change history to conform to woke sensibilities instead of sticking to the facts as they were back when the reality was not marred by subsequent events, in this case, the two MEN changing their names to "solidify" their pretense at changing their "sex."

Another example... a NY Times photo from April 1918 shows Charlie Chaplin being hoisted by Douglas Fairbanks in front of a crowd of thousands on Wall Street. It's a famous photo that depicts when Chaplin was promoting the purchase of Liberty Bonds DURING World War I. The text blithely claims this photo is from AFTER WWI, and goes at length bloviating how Chaplin’s "Tramp" character appearing on Wall Street is an epic demonstration that "juxtaposed" the economic disparities between the financiers and the returning veterans of the post-WWI era… even as the war was still raging when this picture was taken! They were horribly wrong, but went on and on with such pompous and pretentious certainty, that many may take it as fact. See and read for yourself at this link:

Chaplin Photo, pitching Liberty Bonds

(I will try to post some other glaring examples as I come across them. Look in the comments section.)

Third, the effusive overuse of adjectives, beyond what normal experienced human authors, who have learned better, tend to avoid in their writing. And said with a preening as if to show how "humanly" they see things! But it all comes across like a high-schooler trying too hard to come across like an older, more seasoned writer, and it backfires.

(It reminds me of how Data, in Star Trek: The Next Generation, normally speaks without using contractions, i.e., always saying "That is correct" instead of "That's correct" the way most persons would. It's a noticeable affectation that separates him as an android from the rest of the organic crew. When some outside influence or event has him USING contractions, it's so noticeable that it is jarring to the ear and it's the first thing other crew members notice.)

Fourth, there is also something patronizing and smug in the delivery that is harder to pinpoint other than how if "feels" to someone used to reading the works of thousands of humans over the decades. This is where the Turing Test falls short; on the one hand these texts are like something that a human might create, but something is just enough off that it comes across as robotic or fake. Just like some of the Ai images I've seen that can't get hands just right, or have a plastic look that falls short of being authentic.

MOST readers may not notice it, which is what I guess Meta is counting on, but experienced readers, particularly editors, will notice that there's something odd in the overall feel of the texts that, particularly when you see a half-dozen examples on FartBook in one day, seem too much like each other, and not like what any human writer would come up with.

Fifth, there's a level of hyperbole and grandiosity that intrudes, where a more moderate reflection would be appropriate. Each example I read usually repeats and concludes with some vast profound evaluation that sounds more like the epitaph meant for a mausoleum than a sober analysis. It's pretentious and like a movie reviewer trying too hard to ensure that some part of his review will be used in the movie's ads.

Sixth, nearly every example I've come across includes the words "legacy," "enduring," "testament," "legendary," and "iconic," (or most of them) usually in the first and last paragraphs. As if it should be chiseled in stone, always in the same smug and patronizing style.

I am also seeing various articles online that have the same plastic feel. And as the programming improves, and learns to tone it down, it will ever approach "normal" human writing, ensuring that it will eventually replace much of it, and the human writers, who will see their profession evaporate to an increasing degree.

Orwell is spinning in his grave.