AI Analysis
Machine-generated analysis of the post above on 2026-09-09. Not written by the author of the post.
- Post timing (~10:47 PM Eastern) falls in the late-night window associated with more personal posting activity
- No written caption or text is present, so no stylometric evidence (typos, grammar, tangents) exists to support or contradict personal authorship
- Sharing a pre-made video requires no drafting effort, making it low-cost content for staff to queue as well
- No location-change event in the surrounding week suggests a stable Eastern-timezone base for the conversion
Strongest facet: not assessable (no original text)
Primary drive: power
Trigger: Maintenance (Longstanding Big Tech/media censorship grievance narrative; no specific proximate injury identified in the surrounding week's events)
No contradictions with other posts detected yet.
The day started in the middle of a late-night posting binge before his trip to China. In under an hour he pushed out a stream of video clips about Obama, the Russia investigation and the 2020 election, many of them reposts of his own earlier material. After a short night he went after the New York T...
Analysis: Truth Social Post ts_116559318573932949 (2026-05-12)
Content note: This post carries no original text from the subject. The entire payload is a shared video — an interview/testimony segment featuring Zach Vorhies, a self-identified former Google engineer and 2019 whistleblower, discussing Google's "Machine Learning Fairness" program, YouTube content-labeling practices, and alleged politically skewed news-ranking scores. All propositional content in the transcript belongs to the video's speaker, not to the subject. Per framing instructions, the analyzable behavior here is selection and amplification — what he chose to put in front of his audience — not authorship of the words themselves. Analysis below is accordingly thin on trait/style inference and focused on what the choice of content signals.
Level 1 — Dispositional Traits
No stylometric evidence available (no original text). Cannot assess facet-level trait expression from this post. The choice to amplify anti-Big-Tech content is weakly consistent with a longstanding openness-to-experience profile favoring counter-establishment narratives, but this is inference from selection, not from language production, and should be weighted accordingly (low confidence).
Level 2 — Characteristic Adaptations
The clip reinforces a persistent schema in which technology/media institutions (Google, mainstream media, "the establishment") are cast as coordinated, hidden adversaries actively manipulating public perception and elections against him personally. Amplifying this specific whistleblower narrative — which explicitly centers alleged algorithmic targeting of "his fight with Comey" — serves an agency/control motive: reasserting narrative control over a domain (information/media ecosystems) framed as rigged against him. No communion-motive content present.
Level 3 — Narrative Identity
By selecting this clip, the subject implicitly recruits its protagonist (the whistleblower-as-truth-teller) into his own ongoing narrative: a persecuted figure whose adversaries used covert institutional power to disadvantage him, now being vindicated by an insider's disclosure. This is a redemption-adjacent structure (suppressed truth eventually surfaces) grafted onto a broader contamination narrative in which "big tech" corrupted a previously fair system (elections, information access). Contrasting other: Google/YouTube, "mainstream media," implicitly the 2020 electoral apparatus. The subject casts himself, by association, as the vindicated victim/whistleblower-adjacent hero.
Level 4 — Clinical Indicators
Malignant narcissism: Insufficient direct evidence from this post alone (no first-person grandiosity, entitlement, or contempt language present in his own words) to score narcissistic/antisocial/sadistic features. The paranoid-features axis is the most relevant: selecting and amplifying content alleging a large, coordinated, secret campaign specifically targeting him is consistent with a persecutory interpretive frame, though the underlying claims (2019 Project Veritas leak materials) are not novel and have circulated publicly for years — this reads more as reinforcement of an established belief than an acute reactive distortion.
Trigger: Best characterized as maintenance — reinforcing a long-held grievance narrative (tech/media censorship conspiring against him) rather than responding to a specific proximate injury. None of the pre-researched events in the surrounding week (Iran strikes, tariff ruling, Ukraine ceasefire) directly connect to Google/YouTube censorship, so this does not appear reactive to an immediate narcissistic injury; it reads as evergreen grievance content.
Narcissistic state: Not directly assessable from the subject's own language (none present). Indirectly, by proxy through the amplified narrative, the framing is persecutory/victimized (grandiose and vulnerable elements both present in the source material — vindicated hero and wronged party).
Defense mechanisms: No first-person text to assess Vaillant-hierarchy defenses directly. Behaviorally, selecting content that externalizes blame for past electoral/media disadvantage onto a hidden institutional conspiracy is consistent with projection at the level of narrative choice (attributing negative outcomes to external manipulation), but this inference rests on curation behavior, not on his own statements, and should be held at low-to-moderate confidence.
Rhetorical & Propaganda Techniques (in the amplified content)
The video itself (not the subject's words) deploys: appeal to insider authority ("I was a Google engineer... whistleblower"), conspiracy-framing ("sounds like something out of a conspiracy theory, but it's real"), quantified-authority rhetoric (specific numeric "trustworthiness" scores), us-vs-them framing (patriotic truth-tellers vs. corporate manipulators), and implicit whataboutism (comparing Russia Today's score favorably to a domestic conservative outlet to imply reverse-discrimination). By amplifying rather than commenting on this content, the subject lends it his platform's reach and implicit endorsement without taking on direct fact-check liability for any specific claim.
Danger Assessment
No eliminationist language, no dehumanization, no calls to mobilization or violence, no identification of individual targets for action. Danger level: none. The content is a media-literacy/institutional-trust grievance, not a threat-relevant post.
Gaslighting / Reality Distortion
Not directly applicable — the subject makes no first-person factual assertions to evaluate for DARVO or revisionism. The amplified content itself makes contested claims about Google's internal scoring systems; some elements (Project Veritas leak, existence of "Machine Learning Fairness" program) are documented as having been publicly disclosed in 2019, though Google disputed Vorhies's characterization of intent and scope at the time. This dispute is a matter for the video's claims, not the subject's own assertions, and per framing instructions is not entered as a fact-check attributed to him.
Authorship Attribution
No written text exists for stylometric analysis. The only available evidence is timing: 2026-05-12T02:47:35 UTC converts to approximately 10:47 PM Eastern time on May 11 (assuming the subject's default Washington DC/Eastern timezone base, consistent with recent White House-based activity in the surrounding events list — no travel event indicates a location change). This falls within the late-night window historically associated with more personally-directed posting activity rather than staff-scheduled business-hours content. However, sharing a video requires minimal effort (no drafting, no typing) and staff frequently queue or repost supportive media content at any hour, so timing alone is weak evidence for personal selection versus staff curation. Confidence is capped at low, and the attribution question here concerns who chose to share this clip, not who produced its content — the speaker in the video is unambiguously a third party (Zach Vorhies), and no part of the transcript should be read as the subject's own statement.
#
Post from Truth Social
Video transcript 8:39
I was a Google engineer for eight and a half years and a whistleblower against Google in 2019, risking my career and livelihood to warn the public about Google's plans to meddle in the 2020 election using a program called Machine Learning Fairness. Grant, we had a unanimous vote. Every Republican and every Democrat authorized me as the chair to issue these subpoenas. If you've used Google Earth or YouTube for game consoles, you're familiar with my work. Let's talk about Google's censorship AI system called Machine Learning Fairness. I discovered this program while searching for information about Project Dragonfly, probably a fake project meant to misdirect the public. Make no mistake, Machine Learning Fairness is and has always been the real censorship program and it is massive. The goal was to "program the public to align with Google's corporate values." Those are their words. But it's wrong for anyone to have this kind of power to manipulate American elections. And he's right. And in Texas, we're not going to stand for it. We're going to shine a light on what they're doing so that everybody can see. Step one, training data are collected and classified. Step two, algorithms are programmed. Step three, media are filtered, ranked, aggregated or generated. And step four, people like us are programmed. That's a direct quote from their slides. It wasn't just in one slide, it was littered throughout the company. This process was repeated in a cycle with step four feeding back into step one. This sounds like something out of a conspiracy theory, but it's real. Google rewrote their news algorithms specifically trained on mainstream media stories targeting Trump, such as his fight with Comey. Systems like Realtime Events, Realtime Boost, and Hive Mind assigned higher amplification scores to stories related to targeting Trump. Google's internal documents revealed their stance on, quote, algorithmic unfairness. They stated that even factually accurate representations could be considered algorithmically unfair and removed. And let me just quote them. For example, imagine a Google image query for CEOs shows predominantly men, even if it were a factually accurate representation of the world, it would be algorithmic unfairness. In some cases, it may be appropriate to take no action if the system accurately reflects current reality. While in other cases, it may be desirable to consider how we might help society reach a more fair and equitable state via either product intervention or broader corporate social responsibility efforts, end quote. Let's talk about news ranking. News sites that supported Democrats were ranked with the highest quality and trustworthiness. Here's a partial list. Wall Street Journal 8.53, CBS News 6.57, CNN 6.0, Fox News 5.2, Russia Today 4.5, Fox News 5.57, The Young Turks 2.53, Alex Jones Network -1.56. At the bottom of the list is Next News Network at -3.35. Let me reiterate. The foreign propaganda outlet Russia Today was scored seven points higher than the Next News Network, one of the top YouTube news networks in America. Since my disclosure, Next News Network has been permanently demonetized and will soon be bankrupted. These sound more like the actions we'd expect from a Russian propaganda complex. But shockingly, it's the actions of Google. At YouTube, I observed 40 different AIs classifying videos under various labels. Influencers like Dave Rubin were labeled as right-wing news outlets with fake news labels. The justification was always "user safety," but in 2020, this system of censorship went so far as to ban mentions of vitamin C, turmeric, and vitamin D during the pandemic. Anything that is medically unsubstantiated to people saying, "Take vitamin C. Take turmeric. Those will cure you." Those are the examples of things that would be a violation of our policy. Anything that would go against World Health Organization recommendations would be a violation of our policy. And so "remove" is another really important part of our policy. So you're not just putting the truth next to the lie. You're taking the lie down. That's a pretty aggressive approach. Despite public statements to Congress, Google has many blacklists. Some literally named "blacklist.txt." For example, the Eighth Amendment of the Constitution of Ireland, shockingly, was blacklisted from YouTube. Public needs to know what's going on. What recommendations would you have for this legislature to take action to address the issues you address? My recommendation is subpoenas would be filed against Google for any documents that have the name "blacklist.txt" related to Google, YouTube, and their news search corpuses. I would also ask the use of PINA, their documents related to rankings of media outlets. Machine learning fairness is so huge, I don't even know how you would even swallow that. That is going to be such a huge bombshell that it just needs to get into the public disclosure. It's those three things, machine learning fairness, blacklists, and their media rankings. Thank you. Specifically for you, what was the impact of your disclosure? One of the major stories of 2019, the biggest disclosure by Project Veritas, the biggest story they've ever had. We had two disclosures, one in June when I was anonymous. In August, I came out publicly after being raided by the FBI, including the bomb squad, First Department, and SFPD with a standoff that ended at gunpoint. It was all due to a wellness check, and I just didn't want to come out, and so it just kept on escalating, and they shut down the streets from 20th and Valencia to 20 seconds. I realized that if I don't come out, they're going to come in, so I came out, the gunpoint surrendered myself. Literally, Pizzagate. If you look at Google Trends, you'll see that fake news trended one day after the Pizzagate released the files. With the Las Vegas Massacre, that allowed them to inject their YouTube blacklist, which didn't exist prior to this event. From there, it was basically covering up rumors that this was a false flag event, and then it went all the way down to, as I noted, the Eighth Amendment to the Constitution of Ireland. The last term that I saw when I disclosed this list was AMLO. Now what does AMLO mean? That is the nickname of the Mexican president at the time, so they started off possibly with noble goals, but at the end, they were censoring constitutional amendments for our allies and presidential candidates in other countries. Well, how important is it that these subpoenas have been approved to go out, and do you know yet what specifically you'll be asking for? Mexican Senate committees have subpoena power. We rarely use it. It's a big deal. We used it a couple of years ago against BlackRock, State Street, those firms through ESG that are taking Texas money and using it against us. We used it then, and we're using it now. It takes a two-thirds vote in the committee to do this because it's a big deal. Grant, we had a unanimous vote. Every Republican and every Democrat authorized me as the chair to issue these subpoenas to get this information because this affects everyone. You had Professor Epstein, it was a great clip you played. He also told us in the hearing, I'm not a Republican, I lean left, but it's wrong for anyone to have this kind of power to manipulate American elections, and he's right, and in Texas, we're not gonna stand for it. We're gonna shine a light on what they're doing so that everybody can see.
Transcribed automatically. Expect errors in names and numbers.