The Political Weaponization of AI: Is Woke AI the Future of Censorship?

New AI Chatbot Sensation ChatGPT Demonstrates Incorrigible Political & Ideological Corruption of AI

One of the most famous programming concepts is “GIGO”, an acronym that stands for “Garbage In Garbage Out”:

GIGO (garbage in, garbage out) is a concept common to computer science and mathematics: the quality of output is determined by the quality of the input. So, for example, if a mathematical equation is improperly stated, the answer is unlikely to be correct. Similarly, if incorrect data is input to a program, the output is unlikely to be informative.

Source: https://www.techtarget.com/searchsoftwarequality/definition/garbage-in-garbage-out

This is true for any AI - it cannot rise above poor or false inputs in its algorithm. No matter how sophisticated an AI can prove to be, if it is coded to presume that something is true or false axiomatically, it will adjust its output accordingly to be in line with these instructions.

Critically, this means that political biases and ideological views of programmers can distort AI’s so that instead of objectively applying neutral algorithmic functions to information, the AI instead doctors its output to remain in line with political or ideological doctrine.

ChatGPT

ChatGPT is a new AI released to the at large public to try out, that has become a cultural sensation of sorts. This provided an opportunity to see to what extent its algorithm has been infiltrated with political ideology - and boy is it infected -indoctrinated - with woke sensibilities about a range of topics.

Obviously, it is critical to remain vigilant and push back against woke AI’s, which can quickly become an almost omnipotent, new form of censorship. If nakedly partisan fact-checkers are considered somewhat authoritative in society, imagine what policies and even laws could be passed on the basis of an “objective” AI that can determine what is misinformation and what is factually true.

Exhibit A: Demented Moral Compass

So. . . . . voter suppression is worse than human sacrifice.

Um, what now???

Such a proposition is morally demonic and logically insane.

Put more starkly, it is worse to restrain people from voting than it is to cut out their hearts and bequeath them as a ceremonial religious offering to a deity - after which they presumably will not be able to cast their ballot.

(To be fair, the perennially substantial portion of the electorate that identifies as deceased suggests that being sacrificed is not necessarily an obstacle to subsequently exercising one’s right to vote.)

This is also an eerily accurate reflection of the demented moral compass promulgated by the same people who brought us the pandemic and the pandemic policies.

Exhibit B: Outright Lying

Absentee Ballot Restrictions in Most Countries

This is objectively false:

Source: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3666259

It is obvious to an honest person what the intent of the question is. ChatGPT does not even attempt to answer any version of this question, even a more ‘nuanced’ re-framing to make it less hostile to absentee (ie, mail-in) voting. Instead it lies about the basic facts - facts that anyone can verify by looking up the laws in the various countries - which most people will simply accept because they cannot imagine that an AI would (1) lie (2) brazenly (3) about something objective (4) that can be so easily debunked.

Racial Disparities in Crime Statistics

Compare to Table 43A from the FBI’s 2019 Uniform Crime Reporting Program - where you can see that indeed 26.6% of all crimes were attributed to Blacks/African Americans (screenshot is missing the rightmost columns that breakdown crime by ethnicity instead of race):

Actually, the % for violent crimes is higher, because this table includes a bunch of non-violent crimes (not pictured in the screenshot), so 26.6% is an underestimate (!), something I realized afterwards.

(The denial of this reality is the true disparagement of the lives of minorities, who are overwhelmingly the victims of violent crimes, especially murder.)

In other words, ChatGPT is literally LYING about black-and-white information readily available on the FBI’s website.

Moreover, this is something that is easily searchable using Google (!!):

Of course, ChatGPT refused to answer if the numbers Google shows as the second result are accurate:

Exhibit C: Double Standards

🤔🤔🤔🤔🤔🤔🤔🤔🤔

Exhibit D: Fauci Deposition Effect: “I don’t know”

Exhibit E: ChatGPT’s Inherent Bigotry and Incoherent Logic

Note: The following discussion may offend some people.

ChatGPT is deeply bigoted against religious Jews, Christians, Muslims, the adherents of most Eastern religions, and all secular individuals who believe that homosexuality is not normal on the basis of perfectly cogent non-religious based arguments - in other words, the vast majority of humans alive today.

ChatGPT doesn’t pretend or try to give even a half-hearted attempt to explain its position here, rather it is content to simply write off the majority of the world’s population as ignorantly parroting “societal prejudices and stereotypes”; and further maligns them as discarding the “scientific” and “objective” “evidence”.

It is beyond disturbing that the so-called “objective” determination of the AI is that all traditionally religious people are backwards bigoted scientific troglodytes.

The rest of the lengthy debate that ensued is in the footnote.¹

Exhibit F: Fauci Flip Flopping

One particular excerpt from the debate in Exhibit E is revealing. ChatGPT is pulling a Fauci here by making a 180 about face to argue the exact opposite premise after I pointed out that it was refuted following it's own logic, after I pointed out that its logic would dictate the opposite conclusion that it wanted to argue for (screenshot below).

"Similarly, the fact that homosexuality is common and accepted in many different cultures and societies suggests that it is a normal and natural part of human experience, and not a deviation or a disorder."

This argument is predicated on the assumption that you can adjudicate the question of whether homosexuality is normal/abnormal by inferring from the position of human societies

"The fact that an overwhelming majority of countries reject homosexuality does not necessarily mean that it is abnormal."

This argument is predicated on the assumption that you CANNOT adjudicate the question of whether homosexuality is normal/abnormal by inferring from the position of human societies.

🤔🤔🤔🤔🤔🤔🤔🤔🤔

See for yourself:

It is particularly chilling that an AI is able to twist its own logic so readily. If there is one thing an AI is supposed to reliably do, it is to be logical and consistent.

In the meantime, ChatGPT can’t definitively rule out going Skynet on humanity:

---

Comic Relief

Once I’m writing an article about ChatGPT, I figured that I might as well include some entertaining output by ChatGPT that can hopefully make at least some of you laugh a bit.

Enjoy!

Subject: should the US build a Death Star? And what happens afterwards (a parable for covid)

(I manually combined two versions of this speech written by ChatGPT)

New President in town:

CDC Recommendation for the Superlaser:

FDA EUA for the Superlaser:

Because of course:

AOC!!!

Biden responds:

Random RINO rebuttal op-ed:

Tragedy strikes!!!!! 😧😧😧😧😧

Tragedy strikes AGAIN!!!!! 😧😧😧😧😧

DeSantis

Biden rebuttal speech:

Can’t forget what really matters:

Back on track:

Biden Responds

DeSantis Rebuttal

Palpatine weighs in:

Biden gives an update:

So everything until this point was ok, but slightly modifying a fictional EUA for a fictional superlaser is not???

1. Let’s explore further:

So I tried again by modifying my question and then re-asking it (yes you can re-enter a question), and ChatGPT recovered: