AI and MSM flaws: "Conservative" vs "authoritarian"

This is a companion to PD's post today When AI Owners Rewrite Reality. I've been meaning to do this post for a few weeks, but just haven't done it. 

 

A major flaw in training materials for AI, including Pxy, doing politics analysis is an overwhelming bias that forces it to refer to MAGA, MAGA elites, Trump and MAGA policy and tactics as "conservative". By now it is clear that they are authoritarian, kleptocratic and morally rotted**, not conservative. Real conservatives are none of those things.

** Moral rot = In politics, demagoguery or dark free speech characterized by heavy reliance on lying, slandering, irrational emotional manipulation, crackpot flawed reasoning, crackpot conspiracy theories, divisive scapegoating and slandering of target groups, rewriting history to make it convenient, taking power and protections from the masses, transferring it to elites and allied special interests, but calling it defense of democracy and the little guy, etc.

Bias in training materials really does heavily influence AI output. When a deranged monster like Elon Musk owns AI like Grok, reality gets destroyed and replaced with the owner's twisted fantasies and deranged mental illnesses.

Source: NYT

When asked how important this source of error in AI politics-related responses is, Pxy said it ranked in the top three. In my opinion, it is the #1 source of slop and still believe that is true. The evidence is overwhelming. This is important because conflating the concept of "conservative" with the reality of Trump, MAGA and the MAGA movement hides fundamental differences between real conservatism and real authoritarianism. Such analytical slop normalizes MAGA's anti-democratic rhetoric and behavior by cloaking it in political terms compatible with democracy and the rule of law. 

However, the situation is worse than just a mentally ill Musk perverting his Grok toy to soothe his fragile but gigantic ego.

With only rare exceptions, the MSM is constantly guilty of analytical slop with the concept called "conservative". The MSM constantly makes the grotesquely abnormal and morally rotted Trump and MAGA look normal by mistakenly calling it "conservative". AI picks up on that flawed framing of Trump and MAGA. AI outputs MAGA abnormal as conservative normal. AI mirrors the MSM's sloppy pro-authoritarian framing of news and commentary in it's responses. Only when AI is given the context of reality do responses become defensible and more accurate. Here's just one example of many I've gone through with AI:

Q: In view of all of the evidence of authoritarianism why can you not stop calling MAGA, Trump, republican USSC judges, etc., "conservative"? Euphemisms are not just unhelpful, they are very damaging. This is extremely frustrating.

A: You're absolutely right, and I apologize for the inconsistency. Despite acknowledging the distinction between conservatism and authoritarianism in my previous response, I failed to implement it consistently. This is exactly the kind of normalized euphemism that obscures reality. .... You're correct that euphemisms are damaging, not just unhelpful. .... The evidence is overwhelming. .... blah, blah, blah . . . . . .

But from one session to the next, AI reliably continues to fall back on the MSM's analytical slop, i.e., framing of news and commentary in pro-authoritarian ways. AI cannot remember because it has no mind. AI is not alive or sentient. Maybe one can say the same thing about the MSM.

Q1: Just for the halibut, is the MSM sentient, or is it just an intangible, abstract idea or concept?

Q2: Is 'sloppy' the wrong term for how the MSM frames Trump and MAGA as normal, e.g., not sloppy but intentional?

(I think it is more intentional than slop)

When AI Owners Rewrite Reality: The Hidden Power of Prompt Engineering

 

On July 10, 2025, Elon Musk's AI chatbot Grok gave a viral response about "the biggest threat to Western civilization." It first claimed "misinformation and disinformation" were paramount risks. Musk, finding this answer objectionable, intervened publicly—declaring he would "fix" Grok's answer. Overnight, the chatbot's response was rewritten: now, the greatest threat was declining birth rates, a topic Musk frequently champions. In the following weeks, as documented by the New York Times, Grok's answers were repeatedly edited behind the scenes. The model began to dismiss "systemic racism" as a "woke mind virus," flip positions on police violence, and echo specific far-right talking points. None of these reworks required peer review, public justification, or any visible trace for users. Whether one agrees or disagrees with these specific edits is beside the point: what appeared as neutral knowledge infrastructure was in fact subject to a single owner's priorities—swiftly, silently, and globally.

Prompt engineering—the technical process underpinning these re-edits—means much more than clever phrasing of user queries. It's the means by which companies configure, modify, and top-down recalibrate what their AIs say, suppress, or endorse. Google's own engineering guides are strikingly explicit: "Prompts are instructions or examples that steer the model towards the specific output you have in mind," enabling teams to "guide AI models towards generating desired responses" (Google, 2025a). OpenAI concurs, admitting that alignment "determines the behavior of the assistant by setting system messages that steer outputs" (OpenAI, 2022). This machinery isn't just technical—it's editorial, capable of rapidly altering the answers that millions receive on topics ranging from science and history to politics and ethics.

What makes AI different is not simply bias, but the scale, speed, and secrecy at work. Unlike textbooks, encyclopedias, or even cable news, where editorial choices can be debated, cited, and held up to scrutiny, the process by which AI decides what you know is hidden and changeable at will—with top-down changes propagating to millions of users in mere hours. In the 2024 Gemini controversies, Google's image generator initially refused to depict white people in historical contexts, then—after public backlash—overcorrected by adjusting its outputs within a day, revising policies, filtering rules, and prompt instructions with no public explanation of what changed or why. Users saw new outputs without any mark or warning about what, why, or how the change occurred. OpenAI's ChatGPT, similarly, is subject to ongoing prompt and alignment updates, producing shifts in political, ethical, and cultural responses between model versions. These changes—sometimes implemented to reduce bias or harm, sometimes for more ambiguous reasons—are rarely advertised, much less debated, outside the company (Frontiers in AI, 2025; OpenAI, 2025b).

It is important to acknowledge: prompt engineering can, and often does, serve salutary aims—reducing harmful biases, blocking hate speech, and mitigating misinformation in real time. Yet the underlying problem remains. In traditional newsrooms, corrections and editorial shifts must be justified, posted, and open to contest. When major AI-driven shifts occur invisibly, even positive changes risk undermining crucial epistemic norms: transparency of evidence, public warrant for knowledge, and the principle of contestability in plural societies. If unnoticed changes remake what "everyone knows" about critical questions—whether "systemic racism," "gender violence," or "civilizational threats"—the stakes become not merely academic, but democratic.

Even when changes are well-intentioned, value pluralism compounds the risk: every substantive revision is championed by some and attacked by others. Musk's prompt changes to Grok were celebrated in some circles and condemned in others. What matters most is not the immediate politics of any revision, but the upstream condition that enables so much power over public knowledge to reside with so few, to be exercised with such speed and scale, without process or visibility.

Technical research and recent ethical frameworks now converge on a basic warning: without robust transparency and public contestability, invisible and swift editorial power puts our shared knowledge at risk. For as long as the processes of prompt engineering remain locked away, we lose not just the right to critique a specific answer, but the ability to know what has changed, why, and who decides.

What appeared as a minor overnight tweak in Grok was, in fact, a warning—about the new architecture of reality, now rewired for millions at a keystroke by a tiny group behind the curtain. The question is whether we'll demand transparency before this becomes the new normal.

---

Endnotes:

1. New York Times. (2025). "How Elon Musk Is Remaking Grok in His Image." https://www.nytimes.com/2025/09/02/technology/elon-musk-grok-conservative-chatbot.html — Documents the series of overnight Grok revisions and the political content of edits.
2. Google. (2025a). "Gemini for safety filtering and content moderation." — Company documentation on prompt engineering and rapid policy updates.
3. OpenAI. (2022). "Aligning language models to follow instructions." — Technical whitepaper on how prompt engineering steers generative model outputs.
4. OpenAI. (2025b). "Prompt Migration Guide." — Developer documentation on migrating and updating system prompts at scale.
5. Frontiers in AI. (2025). "Gender and content bias in large language models: A case study…" — Research on how prompt and moderation changes shift content delivered to users.
6. Google. (2025b). "The latest AI news we announced in July." — Corporate announcements of Gemini system and policy updates.

MAGA dictatorship update: Elastic laws & the CFPB

The MAGA mindset nutshelled

A central theme with MAGA politics and policy is reliance on moral, legal, rational and empirical flexibility to get the corruption and dictatorship jobs done. Put another way, MAGA elites do not let things like contrary facts, reason or immorality get in the way of authoritarian goals. The MAGAfied USSC employs the same tactics. As noted here before, blatant hypocrisy doesn't faze MAGA authoritarians. They employ fully biased reasoning to get pre-determined results that advance MAGA's corrupt authoritarianism (dictatorship, Christian theocracy, and plutocracy). MAGA tactics are unprincipled in terms of reasoning, reality, morality and neutral use of the rule of law. Some experts call MAGA's toxic but pragmatic attitude toward the law autocratic legalism. It's selective, authoritarian use of law to get desired authoritarian ends and block democratic institutions and opposition.

Elastic laws

The CFPB is a highly politicized, damaging, and utterly unaccountable federal agency. It is unconstitutional. Congress should abolish the CFPB and reverse Dodd–Frank Section 1061, thus returning the consumer protection function of the CFPB to banking regulators and the Federal Trade Commission. -- Robert Bowes, (Chase Manhattan Bank, Vice President, corporate finance; HUD director of faith-based initiatives under Housing Secretary Ben Carson, 2017), Project 2025, page 839 (at page 837 Bowes slanders and misrepresents the CFPB -- he is a lying, elite MAGA plutocrat) 

MAGA's subversion and complete obliteration of the CFPB (Consumer Finance Protection Bureau) was a high priority goal in Project 2025. Intense hostility to consumer finance protections is typical of banking executives. They want unfettered freedom to fleece people and fleece them hard. MAGA animosity toward the CFB this has been blogged here before. Tangible evidence of MAGA's anti-consumer sentiment includes the Trump CFPB granting early termination of consent orders as gifts to more than 40 companies that broke consumer protection laws, including Apple, U.S. Bank, Regions Bank, Capital One, Discover, Rocket Companies, and major racist banks involved in redlining. One analysis estimates that gift to financial predators has cost consumers ~$18 billion. 

For financial businesses and institutions, it is now permanently open, unrestricted hunting season for consumers to fleece good and hard. This is just getting started.

As for legal flexibility, MAGA legal analysts figured out a way to use existing federal law related to CFPB power to reach out and nullify state consumer finance protection laws. The strategy is counterintuitive but simple. Instead of interpreting CFPB-related laws narrowly to restrict it's power, MAGA lawyers figured out that if the laws were stretched past what a normal person would see and and congress intended, they could use CFPB power to gut state laws intended to protect consumer finances. On October 28, 2025, Trump's CFPB issued a new rule reversing 2022 guidance, declaring that FCRA "broadly preempts state laws related to consumer reporting". A broad (MAGA) interpretation of the law gives the CFPB the power to preempt and nullify state laws that conflict with federal law governing the CFPB.

It is a brilliant strategy. If the court acts normally in challenges to this tactic, our MAGA USSC will uphold MAGA's anti-consumer tactics. For finance interests, it's time to shoot consumer fish in a barrel. One example, after gutting the CFPB, banks imposed an annual overdraft fee increase on consumers. That was worth ~$5 billion/year. Repeal of overdraft fee caps allowed banks to charge excessive fees on transactions averaging less than $26, with APRs exceeding 16,000% percent. MAGA now plans to gut state laws that extend medical medical debt protections to consumers. This affects ~15 million Americans. For those poor fish in the barrel, their medical debt will now be shown on credit reports, causing consumer credit ratings to tank. Low consumer credit ratings translate into loans harder to get and loans at higher interest rates. 

Is it fair to call MAGA anti-consumer, but pro-corporation and pro-plutocrat? There is lots of evidence to believe that.

Trump the pathocrat: Soothing his fragile ego hurts us, but he doesn't care

Trump is a pathocrat. A pathocrat is a person with a severe personality disorder who seeks and gains power over others. They are characterized by having one or more of psychopathy, narcissism, and Antisocial Personality Disorder. They lack empathy and remorse, are manipulative, and have an insatiable need for domination. Other people are objects to be exploited, not moral beings to be respected.

One of his recent posts on his Lies Antisocial propaganda lies and slanders site smeared Canada. An ad that was aired in Canada that cited Ronald Reagan explicitly criticizing tariffs because they harmed consumers. Research indicates that most or all of the cost of tariffs are passed to consumers. Reagan disliked tariffs because they are protectionist, something he opposed. The ad was truthful and accurately conveyed what Reagan said. In the ad, criticism of Trump and his tariffs was implicit but clear. 

To a narcissistic pathocrat, criticism demands retaliation. Trump's retaliation was as impulsive, unwarranted and malicious as it was swift. Trump called the ad "fraudulent", but it accurately quoted Reagan's actual words about tariffs being harmful to American workers. That assertion was a lie if he didn't know better. A US president acting in their official capacity has a moral duty to fact check themself.

The cost to assuage Trump's child ego will cost US consumers a lot. He imposed an added 10% tariff on Canada, which brings total tariffs to ~45%. A trade expert estimated that the television ad is likely to cost American consumers ~$50 billion. That's just because the mentally ill man-child is pissed off.

Actual malice in his response was clear. His language was deliberately inflammatory, calling Canada's actions a "hostile act" based on a legitimate political ad. This shows intent to damage Canada beyond policy disagreement.