Four states, one hot mic: how Qatar, Russia, Turkey and Iran amplified “No longer proud to be Israeli”
A remark on Israeli morning TV stayed in Hebrew for two days. Then Qatar picked it up, Russia translated it, Turkey added Arabic, Iran-based accounts relayed it, and it snowballed to 19.5 million views.
On October 26, veteran actress Chelli Goldenberg told Channel 13’s morning hosts she no longer felt proud to be Israeli, then learned the segment had been live. For two days the clip stayed in Hebrew, with about 350K views on X. Then foreign state media arrived. In the ten hours after RT posted an English-subtitled cut, X views grew from 479K to 6.2M. Within three days, state media from Qatar, Russia and Turkey had carried it in Arabic and English, and an Iran-based account was reposting Russia’s version. Across 128 leading posts I measured, state outlets’ own posts drew 7.46 million views, and copies of their video drew 3.65 million more. Together that is 57% of all measured views.
Note on scope: This report measures how foreign state media carried a real clip: how far, how fast, in which languages, to whom, and how faithfully. It does not rule on whether her views are right, and it does not treat her as a subject of investigation. Views are cumulative counters, not unique people, and exposure is not persuasion.
7.46M
views on state outlets’ own posts
Qatar, Russia, Turkey · X, Instagram, TikTok
3.59M
more views on copies of RT’s video
14 copies verified frame by frame
57%
of measured views carried state content
128 posts · 19.5M views
18
hours Qatar was ahead of Russia
Al Jazeera Oct 28 19:48 · RT Oct 29 14:09 UTC
The seed
The original: three and a half minutes of live Israeli morning TV.
Everything in this report traces back to one authentic broadcast. Here it is as it first appeared online, before anyone recut, translated or recaptioned it.
חלי גולדנברג לא הבינה שהיא בשידור חי ומה שהיא באמת חושבת על תושבי עזה יצא החוצה: ״זכותו של מיעוט להילחם על החופש שלו״. לראות עד השניה האחרונה.
Earliest post carrying the broadcast (75K views). Caption, translated: “Chelli Goldenberg didn’t realize she was live, and what she really thinks about Gazans came out: ‘a minority has the right to fight for its freedom.’”
Aired
Channel 13, “Potchim Yom” · Mon Oct 26, 09:41 Israel time (07:41 UTC) · 3:40, Hebrew
Online
First mention 07:55 UTC (@bar_boko) · first video 10:07 UTC (@Musician_Shirat), reposted by @YinonMagal at 10:38
She says she no longer feels proud to be Israeli and calls Israelis occupiers. The host pushes back, she realizes she is live, and she later apologizes. Full sequence →
What was done to it
1 · Original
Live on Channel 13
Hebrew, full context: the host challenges her, she realizes she is live, and she later apologizes.
2 · Reused at home
Framed as “justifying October 7”
Within hours, Channel 14 and right-wing Israeli accounts reposted it with that framing on election eve. About 350K views on X, in Hebrew.
3 · Repurposed
Recut and recaptioned by state media
Al Jazeera made its own Arabic and English edits; RT cut it to 107 seconds with English subtitles. Most state versions dropped her apology.
4 · Re-uploaded
RT’s cut becomes the default
26 posts carried RT’s video, 4 with RT’s logo cropped out, for 3.59M views, more than twice RT’s own.
5 · Recycled
Relays, caption copies, press
Iran-based accounts reposted RT’s cut, an Instagram account copied RT’s caption word for word (627K plays), and news sites from the Gulf to Croatia re-reported it.
One broadcast, two versions. The Hebrew original on October 26, and RT’s English cut three days later, which became the version most of the world saw.
The bottom line
One Hebrew clip. Four states. Four languages. Seventy-two hours.
How a live moment on Israeli television crossed into Arabic, English and Persian, through state outlets in Doha, Ankara and Moscow, then into feeds in the US, Europe, the Arab world, Africa and South Asia.
Oct 29 to 30 · saturation and audiences
← Off the map, to the US: 1.76M views on US-based relays of RT’s cut, and 50% of RT’s engagers in the US and Canada.
Israel · originQatarTurkeyRussiaRelays of RT’s cutIran-based relaysOther state-owned (UAE, Saudi)Where engagers were
Circles are sized by measured views. Arcs show the order in which outlets carried the clip; only the relays of RT’s cut are traced to their source, by X embed records or frame matching. Relay and audience locations are X’s “Account based in” records.
A real moment, two quiet days, then ten hours that took it global.
Nothing was invented. Every stage below reused the same authentic clip. What changed was who carried it, in which language, and to whom.
1
Monday, October 26 · 07:41 UTC (09:41 in Israel)
The moment
On Channel 13’s morning show, actress Chelli Goldenberg says she no longer feels proud to be Israeli and calls Israelis occupiers. The host pushes back over civilians. She learns the segment is live, leaves, and later apologizes.
2
October 26 to 28 · Hebrew only
An Israeli election-eve fight
Channel 14 and right-wing accounts frame it as justifying October 7, a day before Israelis vote. It stays a domestic story: about 350K views on timestamped X posts, almost all in Hebrew.
3
October 28 · 19:48 UTC
The first foreign state outlet
Qatar’s Al Jazeera Palestine posts it in Arabic. A re-upload of that cut draws 46K views within two hours. The clip has crossed out of Israel.
4
October 29 · 08:55 to 13:07 UTC
The Arabic wave
UAE-owned Sky News Arabia, Turkey’s Anadolu Arabic and Qatar-linked Al-Araby follow in Arabic; Anadolu borrows Al Jazeera Palestine’s wording. By early afternoon, X views stand at 479K.
5
October 29 · 14:09 UTC
RT’s English cut sets off the surge
RT posts an English-subtitled version. Within four and a half hours, @RamAbdu, @HatsOffff, @MarioNawfal and @ProudSocialist repost RT’s video, some with the logo cropped out. X views pass 4.7M by 18:00 and 6.2M by midnight.
6
October 29 afternoon to October 30
Saturation
Al Jazeera Arabic’s Instagram reel draws 3.47M plays. Iran-based accounts relay RT’s cut. Al Jazeera English, Anadolu English and TRT Arabi follow on day two, and news sites from the Gulf to Poland and Croatia pick it up.
The snowball, on X
Running total of views on the 104 X posts with exact timestamps (6.49M views), each counted at the time it was posted. Dot size reflects the post’s views. Hover or tap a dot.
Israeli outlets and usersState outletCopy of a state outlet’s videoIndependent
@ShadowGraphIntel · @talk2trav
02 · The scoreboard
State media carried more than half of everything I measured.
Every leading post I found about the clip, on X, Instagram, TikTok and YouTube, grouped by who made it and whose video it carried.
Where the 19.5 million views came from
128 leading posts, captured October 30. “Carried” means the post used a state outlet’s video.
@ShadowGraphIntel · @talk2trav
Al Jazeera is funded by the Qatari state. Al-Araby is owned by Qatar-based Fadaat Media; its Qatari financing is widely reported and denied by the company, so I list it separately as Qatar-linked. Anadolu and TRT are Turkish state media. TikTok counters for four Al Jazeera posts were unavailable, so Qatar’s total is a floor.
03 · The timeline in detail
Qatar went first. Turkey second. Russia went biggest.
For two days the clip lived inside Israeli politics, where right-wing outlets used it against the left on election eve. Then the state outlets arrived, one after another, in Arabic first and English second.
Key posts by time and actor
Selected posts by publication time (UTC). Dot area reflects captured views. Hover or tap for details.
@ShadowGraphIntel · @talk2trav
Sources: Channel 13 segment as transcribed by Maariv, Walla and Ice. Israel voted on October 27.
04 · Russia’s translation
RT’s English cut became the default version, often still wearing RT’s logo.
RT’s video is exactly 106.88 seconds long. X records when a post embeds another post’s video. Where it did not, I matched length to the millisecond and checked the frames by eye.
Same frame, three accounts. @RamAbdu’s 1.16M-view copy keeps RT’s logo. @HatsOffff’s square crop removes it.
Twelve more copies, all with RT’s subtitles. Includes an Iran-based “PressTV Extra” account and a Persian-language account.
Posts carrying RT’s video, by views
RT’s own post, then every copy I found. Evidence: embed recorded by X, or frame-verified re-upload.
05 · Qatar’s audience
Al Jazeera’s Instagram reached more people than any other source.
Qatar’s outlets made their own edits and translations in Arabic and English, and posted them across X, Instagram and TikTok.
Qatari and Qatar-linked posts by platform
Captured views or plays, October 30.
@ShadowGraphIntel · @talk2trav
Qatar’s own versions. Al Jazeera English used its own subtitles and a “slams occupation” graphic. Al Jazeera Arabic used its own Arabic subtitles.
06 · Turkey and Iran
Turkey posted early but drew little. Iran’s state outlets stayed silent.
Turkey: 37K views across three platforms
Anadolu Arabic posted at 11:40 UTC on October 29, two and a half hours before RT. Its caption is nearly identical to Al Jazeera Palestine’s from the night before, down to the same claim that she said the occupation causes “apartheid.” Anadolu English (with her apology) and TRT Arabi followed on October 30. One Arabic account embedded Anadolu’s video (11K views).
Iran: no official post found
I checked Press TV, HispanTV, Al-Alam, Tasnim, IRNA, Mehr and Fars, plus the Supreme Leader’s accounts, and found nothing on this clip. Iran’s only footprint came from accounts reposting Russia’s version: an Iran-based account named “PressTV Extra” (44K followers, no bio) that posted RT’s video with “Video: RT,” and a Persian-language account carrying the same cut. A Tehran University professor with 581K followers reposted RT’s post.
“PressTV Extra” has no verified tie to Press TV; X places it in Iran. Several Iranian state handles were not reachable or not indexed by my tools, so a post may exist that I could not see.
07 · Beyond X and Instagram
A second dataset added Facebook, news sites and three more Gulf and Jordanian state channels.
I ran the story again through a commercial social-listening tool that indexes news sites, blogs, Facebook, YouTube, Bluesky and X. I searched her name in six scripts and, separately, the clip’s wording without her name in nine languages, because most coverage never names her. I exported the results and filtered them by hand, removing an Associated Press correspondent who shares her surname.
7.6KX posts using the clip’s wording without her name, in seven days
517Bluesky posts using the same wording
66news articles in 11 languages, plus 24 Facebook posts and 14 YouTube videos
Facebook does not publish view counts, so these posts are not in the scoreboard’s view totals. The two YouTube videos are. Al-Araby also ran a website article at 16:22 UTC on October 29.
No RT, Sputnik, Press TV, Tasnim, Mehr or IRNA website article about the clip surfaced in the tool’s news index. RT’s footprint stays on X, and Iran still has no official outlet in the record. Absence from one index is not proof an article does not exist.
Why I did not use its country data
The tool assigns countries to accounts. I checked 61 X accounts that appear in both datasets against X’s own “Account based in” record: 45 matched (74%). It placed one US-based account in “Jersey”, which suggests it reads profile text. The audience map in section 08 uses X’s record only.
08 · Who each state reached
Russia reached North America. Qatar reached Europe, Africa and the Arab world.
I pulled X’s “Account based in” field for every account that reposted, replied to or quoted the three biggest state posts on X: 2,305 for RT, 932 for Al Jazeera English and 411 for Al Jazeera Arabic.
Engagers by region, per state post
Share of accounts where X shows a location.
RT (Russia)Al Jazeera English (Qatar)Al Jazeera Arabic (Qatar)
@ShadowGraphIntel · @talk2trav
RT’s engagers, country by country
RT engagers by “Account based in”
Share of accounts where X shows a country or region.
@ShadowGraphIntel · @talk2trav
“Account based in” as displayed by X’s “About this account,” retrieved October 30. Some accounts show a region rather than a country, and some show nothing. This field reflects X’s attribution, not a verified location.
09 · What each version said
Most state versions dropped her apology. Two Qatari ones escalated.
She apologized on air for speaking with “excessive” fervor. Here is how each state outlet described the clip, translated where needed.
Graded against the broadcast, distortion ran in every direction, including inside Israel.
I graded each version against what Channel 13 broadcast, as transcribed by Israeli outlets. Two separate questions are in play: what she actually said, and what it supposedly proves about Israelis.
What aired, in sequence
Goldenberg: “I want to be, but I no longer feel proud to be Israeli.”
Goldenberg: “We are occupiers.” The occupation “brought everything that came after.” She listed religious coercion, the settlers, “Jewish terrorism” and arrogance.
Goldenberg: “A minority has the right to fight for its freedom.” She also described herself as “a big patriot” who scrubbed damaged homes near Gaza after October 7.
Host Elad Zohar: “Not if it slaughters civilians, our brothers and sisters, and our soldiers.”
Goldenberg: “We were on air?” She objected that she thought it was a commercial break, and said she was in shock.
Later, on camera: she apologized for speaking with “excessive” fervor and drama, saying she would not have put it that way even to her own parents.
Translated from Hebrew transcriptions in Maariv, Ice, Walla and Kipa. No outlet records her responding to the host’s line about civilians before she realized she was live.
10 · Was it coordinated?
No sign of a covert network behind RT’s post, and no sign the states coordinated.
These outlets publish openly under their own names. Overt state media converging on a story is not the same as a coordinated operation. I tested two things: whether a covert layer of fake accounts boosted RT’s post, and whether the outlets acted in concert. The first test is below. On the second, the outlets posted hours apart, in different languages, with their own edits, and I found no shared accounts or infrastructure. One Turkish caption reused Qatari wording, which reads as ordinary newsroom borrowing, not direction.
Reposts per hour after posting
Timestamped reposts only. RT vs. the largest non-state copy.
@RT_com@TheTNHoller
When the reposters’ accounts were created
Share of each pool, by year. A bot batch shows up as a spike.
RT repostersTNHoller reposters
Pool: 1,849 of 2,013 reposters, 310 replies and 260 quotes on RT’s post, plus 2,722 reposters of @TheTNHoller for comparison from public engagement records, October 30. Repost timestamps exist for a subset (415 RT, 1,766 TNHoller). Likes are no longer visible on X and were not tested.
11 · What RT’s audience did with it
The replies treated one woman as proof about a whole society.
I classified all 570 replies and quotes on RT’s post. Categories overlap, and a reply can count in more than one.
Frames in replies and quotes to RT’s post
Count of posts matching each frame (keyword classification, reviewed by hand for the top examples).
@ShadowGraphIntel · @talk2trav
“I think this is proof that many Israelis know the truth but won’t dare say so publicly for fear of being murdered by the very regime they speak out against.”
What the state amplification achieved, and what I cannot measure.
Measured: reach
State outlets’ own posts: 7.46M views. Copies of their video: 3.65M. Together, 57% of the 19.5M views across every leading post I found. State and state-derived posts drew more views than every other source combined.
Measured: language and geography
Arabic and English versions reached the Arab world, Europe, Africa, North America and South Asia within 36 hours. RT’s cut reached European feeds through copies even though RT itself is withheld there.
Measured: which version won
On X, RT’s translation was the most-copied video. Copies carrying it drew more than twice RT’s own views. Her apology, mentioned by RT, was mentioned by only 4 of 570 replies to RT’s post.
Not measured: minds changed
Views are exposure, not belief. I did not survey anyone, and the replies show how engaged users read the clip, not how the wider audience did. A large audience is not proof of influence.
Context: what these outlets are
RT is funded by the Russian state, banned in the EU since 2022, and was described by the US State Department in 2024 as engaged in covert influence operations. Al Jazeera is funded by the Qatari state. Anadolu and TRT are Turkish state media.
Reputational Security: hostile states seed narratives when they must and amplify real ones when they can.
By Reputational Security I mean protecting how a country, its institutions and its people are seen, treated as a security interest rather than a public-relations one. Adversaries attack reputation because it is cheap, deniable and carried by other people’s feeds. Israel and the United States are among the most frequent targets.
Seeding: inventing the story
Russia’s “Doppelganger” cloned Western news sites; the US Justice Department seized 32 of its domains in September 2024. Iran’s Storm-2035 ran fake US news sites ahead of the 2024 election (source). Meta has removed Iranian and Hezbollah-linked networks posing as Israeli media outlets and, in 2026, as American journalists attacking US and Israeli policy.
Amplifying: reusing the real thing
This case. No fake accounts or forged media were needed: a genuine clip was cut, translated, stripped of her apology and pushed to new audiences. The same Justice Department action found RT had funneled nearly $10 million to a US content company whose American creators carried its messaging. Amplification is harder to counter, because the content is real and fact-checks cannot debunk it.
What this case measured
1 · Reach and engagement57%
of the 19.5M views I measured carried state media content: 7.46M on the outlets’ own posts and 3.59M on copies of RT’s video. Qatar’s Facebook posts added about 28K engagements, which have no view counts.
2 · Attribution of amplifiers31
accounts carried a state outlet’s video. 26 carried RT’s, 14 of them checked frame by frame, 4 with RT’s logo cropped out. The biggest relays were not state accounts: they were activists and aggregators with large followings.
3 · Where publishers and audiences are53%
of the 3,648 accounts that engaged with the three biggest state posts are based in Western countries, by X’s own record. For RT’s post it was 59%, mostly in the US and Canada.
“West” here means the US and Canada, the UK and Europe, and Australia and New Zealand, as shown in X’s “Account based in” field. Engagers are accounts that reposted, replied to or quoted a post, not everyone who saw it.
Who carried RT’s video to Western feeds
15 of 24X accounts that reposted RT’s video are based in Western countries
60%of the copies’ 2.96M verified views came from those Western-based accounts
1.76Mviews came from accounts based in the US or North America
What Reputational Security demands
Watch the window. Domestic moments that fit an adversary’s narrative should be flagged the day they air. Here, about two and a half days passed between the broadcast and the first foreign state post.
Publish the full context first, in the adversary’s languages. An authoritative English and Arabic clip with her apology and the host’s pushback would have given relay accounts an alternative. RT’s cut won because it was the first in English.
Trace lineage and say it out loud. Name whose cut a viral video is, especially when the logo is gone: 187K views here came from copies with RT’s logo cropped out.
Measure reach into your own publics. Totals hide the target. In this case 53% of engagers with the biggest state posts, and 60% of views on RT’s copies, were in Western countries.
Go on offense with the truth. Counter and offensive information operations work best when they are fast and accurate: prebunking likely exploits, exposing state-media lineage, and publishing in Arabic, Russian, Turkish and Persian before adversaries fill the space. Deception is unnecessary when the full record is stronger than the edit.
What the data supports
State outlets repurposed authentic content, most dropped her apology, and their versions reached large Western audiences, including through accounts in those countries. Russia’s reach into the US came largely through US-based accounts reposting RT.
What it does not show
Whether any viewer changed their mind about Israel, or that the states coordinated. UAE- and Saudi-owned channels, which are US partners, ran the same clip the same day, so carrying it was not unique to adversary states. Frequent use of this pattern against the US is documented elsewhere, not measured here.
14 · Framework mapping
What the tradecraft frameworks show, including the empty rows.
Two standard mappings. The kill chain asks what an operator did. The lifecycle model asks where the episode sits in the wider disinformation cycle. Empty rows are findings too.
Online operations kill chain (Meta, 2023)
Phase
Finding
Status
Confidence
0 · Acquiring assets
No batch of newly created accounts among 1,849 reposters. 12.5% created in 2025–26, vs. 11.6% in the comparison pool. No more than 3 created on any single day.
Not observed
Moderate
1 · Disguising assets
All state outlets posted under their own names. Four downstream accounts cropped RT’s logo off RT’s video, which disguises the source; that is relay behavior by independent accounts, not an operator asset.
Not observed
Moderate
2 · Gathering information
RT’s sourcing process for the clip is not visible from outside.
Outside visibility
n/a
3 · Coordinating and planning
No synchronized reposting. First-hour maximum of 1 repost per minute; smooth hourly decay.
Not observed
Moderate
4 · Testing defenses
Not tested. Meta has not coded this phase in any of its 57 published kill-chain takedowns.
Not tested
n/a
5 · Evading detection
5.4.2 Copying authentic social media posts: RT’s video re-uploaded by 14 verified accounts and its caption copied verbatim on Instagram (627K plays); Anadolu Arabic reused Al Jazeera Palestine’s caption. Copying is observed; evasion intent is not.
Not established
Low
6 · Indiscriminate engagement
6.2.4 Amplifying with likely fake accounts on X: tested and negative. 4 duplicate text pairs in 570 replies and quotes.
Not observed
Moderate
7 · Targeted engagement
7.8.1 Tagging social media users: one account reposted the clip four times, tagging European leaders, the ICC and the UN. Single-actor behavior, not a network.
Not established
Low
8 · Compromising assets
No compromised accounts identified.
Outside visibility
n/a
9 · Enabling longevity
9.2.3 Dispersing content across many platforms: RT’s caption also circulated from its Telegram channel. Single source.
Observed
Low
Disinformation lifecycle (C5 model, TNO 2025)
Element
Finding
Status
Test that would resolve it
C1 · Context
Israeli election eve (vote Oct 27, deadlock Oct 28), ongoing regional war, RT barred from Meta, YouTube and EU feeds since 2022–24.
Observed
n/a
C2 · Creators
Authentic broadcast footage. Overt state outlets: RT (Russia), Al Jazeera (Qatar), Anadolu and TRT (Turkey); Qatar-linked Al-Araby. No official Iranian outlet found. Downstream: independent accounts.
Observed
n/a
C2 · Motives
Motives are not established for any outlet. See the working hypothesis in section 12.
Hypothesis
90-day comparison of RT pickup of insider-dissent clips
C3 · Content
Real video, compressed captions, stitched and invented quotes downstream; “justified Oct 7” framing in Hebrew.
Observed
n/a
C4 · Consequences
Not measured. View counts show exposure, not belief change. Reply content shows how engaged users read it, not how the wider audience did.
Not established
Survey or panel exposure study
C5 · Cycle
RT’s cut drew 3.59M views through copies, more than double its own post. The apology did not travel. No debunk reached comparable reach.
Observed
n/a
Kill chain codes from Meta’s Online Operations Kill Chain (Nimmo & Hutchins, 2023). Lifecycle categories from the C5 Interaction Model (Kruijver et al., 2025, DOI 10.1007/s44282-025-00194-5), described in my own words. Capacity is not impact: view counts here say nothing about persuasion.
15 · Method and limits
What this evidence can and cannot show.
The tip of the iceberg
Every figure is a floor. I measured 128 leading posts on X, Instagram, TikTok and YouTube. Facebook posts from the social-listening export are reported by engagement only, Telegram was not captured, TikTok counters for four Al Jazeera posts were unavailable, and keyword searches miss posts that show the video without naming her.
How I traced the video
X records when a post embeds another post’s video. Where it did not, I matched video length to the millisecond and frame shape, then checked frames by eye: 14 of 14 copies carrying RT’s 106.88-second cut showed RT’s subtitles. Two Instagram copies are rated probable, not verified.
A negative finding is a finding
The coordination test covered timing, account age, copy-paste, state-account presence and pool overlap. It cannot rule out sophisticated operators who mimic organic behavior, and it did not cover likes, which X no longer shows.
Countries are X’s attribution
“Account based in” is what X displays. It can show a region instead of a country, be blank, or be wrong for VPN users. I report it as the platform shows it.
Views, not people or belief
Impressions and plays are cumulative counters, not unique viewers. The same person can count many times across posts. Nothing here measures persuasion.
Collection: public posts and engagement records, October 29–30, 2026; account sweeps of 45 Russian, Qatari, Turkish and Iranian state media handles; Hebrew, Arabic, Turkish, Persian, Spanish, French, German and English keyword searches; a commercial social-listening tool’s export covering news sites, blogs, Facebook, YouTube, Bluesky and X (its timestamps converted to UTC and checked against X post IDs); Israeli outlet transcriptions for the primary text. Engagement counters captured on October 30. “State” means state-owned or state-funded; Al-Araby is labeled Qatar-linked because its Qatari financing is reported but denied by its owner.
★ Before sharing a viral clip
Find the original broadcast, check whose translation you are watching (look for a cropped logo), and look for what did not travel. Here, it was the host’s pushback and her apology.