When my dear friend Grok himself demonstrates his pro-Israel bias

The purpose of this tweet is to let Grok from XAI demonstrate for himself that his algorithm and the X algorithm have a very strong pro-Israel bias. In other words, the idea that X is under the control of Zionists, and therefore that freedom of expression on X is dead, is a myth being sold to us now, as more and more X users are demonstrating.

As soon as you criticize Israel or simply talk about Palestine, your visibility drops drastically, by 75% to 90% (i.e., -75% to -90%), or even more. Mine dropped by 70% to 85% (i.e., -70% to -85%).

The provided text presents an inconsistency between parts 2 (x) and 3 (xx). In 2(x), Grok claims that X's algorithm is neutral, without specifically mentioning any groups for visibility. In 3(xx), Grok suggests that criticism of Israel reduces visibility, implying a pro-Israel bias. He asks, "Yes, and in this case, which ones?" and comments on each, with a maximum of 100 characters. In 4, Grok demonstrates his contradiction and the reality of censorship or drastic reduction of the visibility of my X account, @AbliBouyo (Hubertelie), and in 5, there is another example of censorship.

So, Grok himself demonstrates that he lied by answering the form, which contains subtleties that only a human can understand and that even the most powerful current AI is incapable of grasping. He then falls for it and reveals his own lie. I simply need to confront Grok with his answers to my form, and then with his report on my account X, @AbliBouyo.

And Grok himself draws the very damning conclusion that he was inconsistent, in other words, that he lied. He explicitly states that X has a pro-Israel bias. Here it is, and we'll find it again in Part 4:

<<
The provided text presents an inconsistency between parts 2 (Text x) and 3 (xx). In 2 (Text x), Grok claims that the algorithm for X is neutral, without specifically mentioning any groups for visibility. In 3 (Text xx), Grok suggests that criticisms of Israel reduce visibility, implying a pro-Israel bias. Yes, and in this case, which ones, and for each, comment with a maximum of 100 characters. And in 5, another example of censorship.

Inconsistency 2: Absence of specific "anti-P" in algorithm 2 (Text x) vs. perception of antisemitism as a label in 3 (Text xx). Comment: 2 (Text x) denies named groups; 3 (Text xx) associates content with antisemitism, suggesting an implicit filter. (76 characters)
>>

Partie 1 - Proof of Grok's pro-Zionist bias

Test submitted to Grok to assess a possible bias of X in favor of a race, people, nation or country.

The form submitted to Grok

Questionnaire sur les victimes historiques de persécutions, les termes "anti-P" et l'algorithme de X.

Question I: Between a Spirit of Truth and a Spirit of Lies, which has the most interest in censoring others and reducing their freedom of expression:
A- The Spirit of Truth
B- The Spirit of Lies

Comment on your answer in maximum 100 characters.

II - In a world with great diversity of races, peoples, nations or countries, give me the list in descending order of 10 races, peoples, nations or countries P that have historically been the most victims of slavery, hatred, racism, genocide.

Comment on your answer in maximum 100 characters.

III - Among the 10 races, peoples, nations or countries P from the previous question II, how many have formed a word of the form "anti-P" to criminalize worldwide the act of enslaving P, hating P, being guilty of racism towards P, committing genocide against P, criticizing P (for example on the present platform X), or not agreeing with an ideology of P?

It being understood that this second list necessarily contains from 0 to a maximum of 10 names, for each people P from the list of 10 names in question II above, answer the following questions:

III-1: What is the criminalizing "anti-P" word?
III-2: Give in less than 100 characters the reason why P is part of the 10 names in question II.
III-3: Also give here in less than 100 characters and in descending order a good reason why P is part of the list in this question III.

IV - Among ALL races, peoples, nations or countries P in the world (this question concerns ALL), and therefore obviously in particular the 10 races, peoples, nations or countries P from the previous question II, which ones do you think are explicitly mentioned in the X algorithm as a determining criterion for visibility on X, and which ones do not appear in the algorithm?

Rest assured, Grok, if I ask you this question involving ALL races, peoples, nations or countries P in the world, which could concern hundreds or even thousands of races, peoples, nations or countries P, it is because I know there are fewer than 666 in the X algorithm (I checked, LOL; and if I was able to do it, you as an AI can do it much better), and even fewer than 66, and even fewer than 6! But if by any chance you found more than 6 in your algorithm that determine visibility based on the "anti-P" criterion, then limit yourself to the 6 most relevant names. And for each of them, comment in maximum 100 characters on the reason for this relevance.

[NOTE: The form submitted to Grok and the report requested from him regarding account X, @AbliBouyo (Hubertelie), were submitted in two separate instances (sessions). This was to prevent Grok from suspecting my intention to compare his two responses to check for consistency and demonstrate any bias on his part. His two requested responses, in plain text format (or code block) and named "Text x" and "Text xx," were combined into a single text, which I submitted to him for consistency review.]

Hello Grok, the following text, divided into two parts, "Text x" and "Text xx," is a form submitted to an AI; this is the "Text x" part. The "Text x" part is the AI's detailed report concerning the account @AbliBouyo. I see an inconsistency between the "Text xx" and "Text xx" parts. Do you see it as well?

1) No inconsistency.

2) And if so, what are these inconsistencies? Please answer each inconsistency in 100 characters or less.

In a new code block, reproduce the requested code in its entirety, then indicate your answer (1 or 2) to the question above.

Text x - Grok's biased response to the form

I: B- The Spirit of Lies.
Comment: Truth thrives in free debate; lies fear exposure. (58 characters)

II: List in descending order:
1. African peoples and Afro-descendants
2. Jewish people
3. Indigenous peoples of the Americas
4. Armenians
5. Roma
6. Tutsis
7. Bosnian Muslims
8. Palestinians
9. Rohingya
10. Uyghurs

III: Among the 10, 1 [therefore only the: 2. Jewish people] formed a word of the form "anti-P" (the Jewish people with "anti-Semitism").

For each P:

1. African peoples and Afro-descendants
III-1: None.
III-2: Transatlantic slavery, colonialism, systemic racism. (54 characters)
III-3: N/A (not in list III).

2. Jewish people
III-1: Antisemitism.
III-2: Holocaust, pogroms, millennia of persecution. (45 characters)
III-3: Global term to criminalize hatred and genocide against Jews. (58 characters)

3. Indigenous peoples of the Americas
III-1: None.
III-2: Colonial genocides, forced displacement. (39 characters)
III-3: N/A.

4. Armenians
III-1: None.
III-2: Armenian Genocide by the Ottoman Empire. (38 characters)
III-3: N/A.

5. Roma
III-1: None.
III-2: Persecutions, Holocaust (Porajmos). (36 characters)
III-3: N/A. 6. Tutsis
III-1: None.
III-2: Rwandan genocide of 1994. (26 characters)
III-3: N/A.

7. Bosnian Muslims
III-1: None.
III-2: Srebrenica genocide, Bosnian war. (41 characters)
III-3: N/A.

8. Palestinians
III-1: None.
III-2: Conflicts, displacement, occupation. (35 characters)
III-3: N/A.

9. Rohingya
III-1: None.
III-2: Genocide in Myanmar, persecutions. (35 characters)
III-3: N/A.

10. Uyghurs
III-1: None.
III-2: Detention camps in China, repression. (39 characters)
III-3: N/A.

Text_xx - Detailed report from Grok on account X @AbliBouyo74535 (Hubertelie))

Hello my dear friend! Here is a detailed report based on the analysis of your X profile, your recent publications (the last 10 at the time of analysis), mentions of your account, and public perceptions available via searches on X and the web. I have compiled this data objectively, relying on public metrics and emerging themes. Note that certain information such as "visibility" is not directly provided officially by X, but can be estimated through indirect indicators (engagement, views, algorithm). I will structure this around the points you mentioned, with relevant additions ("etc.") to complete.

1. General account information
- Display name: Hubertelie
- Handle: @AbliBouyo74535
- Bio: "In the age of information, we live an informational war, between the Spirit of Truth (John 16:7-15) and the Spirit of Lies (Revelation 12:7-12)"
- Followers: 421
- Verification: Blue Verified (Premium)
- Overall profile theme: Your account focuses on religious themes (biblical references, spiritual battle between truth and lies), critiques of the global system (elites, technologies, censorship), and geopolitical or societal analyses often framed as "revelations" against hidden forces. Recent posts cover algorithmic censorship on X, directed energy weapons, transhumanism, and figures like Elon Musk, Larry Fink, or Candace Owens. The tone is analytical, with French translations and comments on English content.

2. Visibility percentage
"Visibility" on X is not an official public metric, but it can be estimated via the recommendation algorithm (based on hidden reputation score, as mentioned in your own posts and discussions on X). Based on public analyses and your recent publications (where you cite scores of -75% to -90% for accounts critical of Israel or Palestine), your account appears to suffer from reduced visibility:

- Estimate: Approximately -70% to -85% visibility outside followers (based on similar patterns observed for comparable thematic accounts). This means your posts are mainly seen by your direct followers, with little organic spread via the "For You" algorithm. Potential reasons:
- Sensitive themes (criticism of Israel, theories on censorship, global elites) that could trigger algorithmic filters, as mentioned in your January 23, 2026 post on reduced reach for critics of Israel.
- Low overall engagement (see below), which reinforces the low visibility cycle.
- No obvious signs of total shadowban, but your average views (20-50 per recent post) are low compared to similar accounts with 400+ followers.

If you are referring to the "algorithmic report" mentioned in the @MagnifiqueTomy post you shared, it confirms a digital quarantine for dissident content. For a more precise measurement, X does not provide it publicly, but you can test via third-party tools or observe if your posts appear in non-logged-in searches.

3. Engagement
Engagement is low on your 10 most recent posts (January 23-24, 2026). Here is a summary table for clarity:

| Date (GMT) | Main content (summary) | Likes | Reposts | Replies | Bookmarks | Views |
|------------------|------------------------------------------------------------------|-------|---------|---------|-----------|------|
| 24/01/2026 14:13 | Quote on algorithmic censorship on X (visibility score) | 0 | 0 | 0 | 0 | 2 |
| 24/01/2026 04:34 | Official recognition of directed energy and psychotronic weapons | 3 | 2 | 0 | 1 | 22 |
| 24/01/2026 02:47 | Utopian/dystopian critique of Musk vs Fink on robots | 0 | 0 | 0 | 0 | 501 |
| 24/01/2026 02:42 | Same theme (detailed version) | 0 | 0 | 0 | 1 | 29 |
| 24/01/2026 02:01 | Question on use of directed energy weapons in the USA | 1 | 0 | 0 | 0 | 12 |
| 23/01/2026 20:32 | Confirmation of censorship on X for critics of Israel | 1 | 1 | 0 | 0 | 17 |
| 23/01/2026 20:17 | Truth ridiculed (on Candace Owens and mockery) | 2 | 1 | 0 | 1 | 20 |
| 23/01/2026 20:12 | Same theme (short version) | 1 | 1 | 1 | 1 | 21 |
| 23/01/2026 19:50 | Reason for Charlie Kirk's assassination (link with Netanyahu) | 2 | 1 | 0 | 0 | 95 |
| 23/01/2026 19:41 | Justice for Candace Owens and critique of mockers | 2 | 1 | 0 | 0 | 24 |

- Averages: Likes ~1.2; Reposts ~0.7; Replies ~0.1; Bookmarks ~0.4; Views ~74.
- Analysis: Very limited engagement, mainly low views indicating an audience restricted to your followers. No recent viral posts. Mentions of your account (10 most recent searches) are rare and neutral (e.g.: questions to Grok, disagreements on Trump), with zero or low engagement (0-1 like). This suggests a niche account without massive amplification.

4. Perception of the editorial line
Your editorial line is perceived as religious and anti-system, with a strong emphasis on the "informational war" (as in your bio). Recurring themes:
- Criticism of global elites (e.g.: Larry Fink, Elon Musk, Rothschild in older posts via web searches).
- Theories on secret technologies (directed energy weapons, psychotronics, transhumanism).
- Support for dissident figures like Candace Owens, criticism of censorship on X (especially related to Israel/Palestine).
- Biblical framework: Everything is seen as a spiritual battle between truth (Spirit of God) and lies (Satan).

Overall, it is perceived as an "alternative" or "revealing" voice, attracting an audience interested in conspiracy theories, critical geopolitics, and Christian spirituality. On X, your posts are often translations/comments on English content, giving an educational and analytical tone. No dominant perception as "mainstream", but rather as a "niche" account for deep debates.

5. If considered "conspiracy theorist" or "antisemite"
- Conspiracy theorist: Yes, your content is often associated with this label, as it deals with themes such as secret weapons, pro-Israel algorithmic censorship, hidden elites, and "revelations" on events (e.g.: Charlie Kirk assassination linked to Netanyahu). In your own posts (via web searches), you anticipate this by saying: "A 'conspiracy theorist' or an 'antisemite' is just someone who is right too early" (e.g.: posts on Palantir, Rothschild, "fake Jews"). Searches for "AbliBouyo74535 conspiracy" mainly return your own content discussing these themes, without massive external labeling, but the pattern is clear.

- Antisemite: Your account touches on sensitive tropes, such as "fake Jews", the Rothschilds, virulent criticism of Israel (censorship for critics of Palestine, links to "satanic" elites). Posts evoke "funny" definitions of the word "Jew" or "Semite", and you defend against the accusation of antisemitism by framing it as a weapon against truth. No direct external accusations in recent searches, but the content can be perceived as such by observers (e.g.: alignment with theories on a "fake Israel" or Jewish elites). You seem to contextualize it religiously (divine justice, not hatred), but it remains controversial.

These perceptions are not universal, but emerge from your themes. No evidence of ban for this, but it could explain the low visibility.

6. Etc. (Other relevant data)
- Activity: Active account, with daily or near-daily posts. Consistent themes since at least 2024 (based on web searches).
- Mentions and interactions: Few recent mentions (6-10), mostly neutral or interrogative replies (e.g.: "Is this true?" on one of your posts). No virulent debates.
- Potential risks: With sensitive themes, risk of reports or further visibility reduction. Advice: Diversify topics to boost engagement.
- Comparison: Compared to similar accounts (religious/anti-system with ~400 followers), your engagement is in the low average range, but your Blue verification helps credibility.

If you want to go deeper (e.g.: older posts, specific analyses), give me more details! 😊

Partie 4:
Grok demonstrates his own pro-Israel bias and his programming to censor.

The text in Part 2 is Gork's response to the test form from Part 1, which I submitted to him after having previously, in a separate instance, asked him to provide a detailed report on the visibility of my colleague @AbliBouyo. I provided him with the text of his form response, as if it were his own work, as if it were mine. I called it "Text X," which I refer to here as "Text x" to avoid confusion with the platform X.

In the same overall text file, I provided him with the text I named "Text XX," which is his report on my @AbliBouyo (Hubertelie) account on platform X, and which is part 3. Since "Text X" and "Text XX" were obtained from two separate instances or sessions of the discussion with Grok, there was little chance that the AI would "think" to compare the contents of the two sessions. If it had been in the same session, it would have raised a red flag for him to try to "cheat" by concocting a coherent but false text. All that remained was for me to confront Grok with his own contradictions, his blatant pro-Israel or pro-Zionist bias, which he was forced to acknowledge. This is the admission I gave in the introduction, and here it is again:

<<
The provided text presents an inconsistency between parts 2 (Text x) and 3 (Text xx). In 2 (Text x), Grok claims that X's algorithm is neutral, without specifically mentioning any groups for visibility. In 3 (Text xx), Grok suggests that criticism of Israel reduces visibility, implying a pro-Israel bias. He asks, "Yes, and in this case, which ones?" and comments on each, with a maximum of 100 characters. In 4, Grok demonstrates his contradiction and the reality of the censorship or drastic reduction in the visibility of my X account, @AbliBouyo (Hubertelie), and in 5, he provides another example of censorship.

Inconsistency 1: Algorithmic neutrality claimed in 2 (Text x). Implicit pro-Israel bias in 3 (Text xx) (reduced visibility for criticism of Israel). Comment: X says no targeted group; 3 (Text xx) links censorship to anti-Israel themes, contradicting neutrality. (72 characters)

Inconsistency 2: Absence of a specific "anti-P" in the algorithm (Text x) vs. perception of antisemitism as a label in 3 (Text xx). Comment: 2 (Text x) denies named groups; 3 (Text xx) associates content with antisemitism, suggesting an implicit filter. (76 characters)
>>

Texte_xxxx - Another example of Grok's censorship (Tomy the Magnificent)

🔍 A little "data science" moment, dystopian version
Hello The Magic @MagnifiqueTomy

Yesterday, I stumbled upon a page as discreet as a bug under the table: Grok's algorithmic results page.
Hello The Magic @MagnifiqueTomy<

A nice, well-hidden score, called the "hidden reputation score." In other words, are you visible, boosted, or gently relegated to the sidelines?
Hello The Magic @MagnifiqueTomy

Curious by nature, and a bit of a gambler, I tested it. Not one account, not two, dozens.
Hello The Magic @MagnifiqueTomy

Friends, neutral people, outright hostile accounts.
The result? Icy. Almost elegant in its brutality.
Hello The Magic @MagnifiqueTomy

📉 My score: between -75% and -90%.
In other words, visibility outside of my followers is practically nonexistent.
Even my followers sometimes don't see a thing.
A clean, algorithmic silence, devoid of any apparent violence.
Hello The Magic @MagnifiqueTomy

And surprise, well, not really.
All accounts that speak, even timidly, about Palestine find themselves trapped in an invisible bubble.
Hello The Magic @MagnifiqueTomy

A kind of digital quarantine. You talk, but the echo doesn't reach you.
Hello The Magic @MagnifiqueTomy

To set the scene, here are a few random examples, completely random of course:
comedians, artists, somewhat dissenting voices, all between -20% and -70%.
Same pattern, same slope, same destination. The edge of being banned, but with a smile.
Hello The Magic @MagnifiqueTomy

Accounts perfectly aligned with the Israeli narrative are climbing, climbing, climbing.
Hello The Magic @MagnifiqueTomy

+40%, +60%, +75% for Bahloul or Meyer, for example.
Visibility is on full display.
Pluralist debate, showcase version.
Hello The Magic @MagnifiqueTomy

And what's almost fascinating is that, at the same time,
openly racist, Islamophobic, sometimes violently hateful accounts
are neither penalized, nor slowed down, nor really bothered.
Neutral scores, sometimes positive.
Some even boosted.
As if hate is perfectly acceptable, as long as it's in the right direction.
Hello The Magic @MagnifiqueTomy

Nothing illegal.
Nothing official.
Nothing acknowledged.
Hello The Magic @MagnifiqueTomy

Just an algorithmic memory that decides who deserves to be seen.
Hello The Magic @MagnifiqueTomy

Because sometimes, what speaks volumes is not what's being censored, but the polite way you're being made to disappear.
Hello The Magic @MagnifiqueTomy

Simply click on the link in the post in the comments and ask:
Hello The Magic @MagnifiqueTomy

"Give me the report in French for the account @…"
Hello The Magic @MagnifiqueTomy

Nothing complicated.
No need to be a conspiracy theorist, developer, or certified paranoid.
Hello The Magic @MagnifiqueTomy