Cyabra Launches Deepfake Detection

Taliban’s takeover of Afghanistan

Identifying fake stories spread on Twitter following the Taliban takeover of Afghanistan.

The Taliban’s takeover of Afghanistan sparked widespread discussion on Twitter. More than 300,000 profiles that took part in the conversation were scanned, and more than 32,000 examined. Of those, 10.3% were deemed inauthentic. Fake stories falsely claiming that the Taliban had raised their flag over the Presidential Palace, and that the journalist Bernie Gores had been killed, were spread by both inauthentic and authentic profiles.

#National Security

Share this report
Composite tweets alleging Taliban flag over Kabul presidential palace, overlaid with grey 'ALTERED' text and red 'FAKE' stamp
Share this report

Fill up the form below and receive the full report directly to your inbox

Related reports

Coordinated Disinformation Targeting Romania’s Election

January 21, 2025

Following the annulment of Romania's election, Cyabra revealed a coordinated campaign backed by Russia aimed at manipulating the country’s electoral crisis.

Cyabra dashboard displaying analysis of 1,038 social media profiles with 137 flagged as inauthentic and 879 authentic, alongside several example X posts labeled “FAKE” that spread false claims about Romania’s presidential elections

#Elections, National Security, Threat Actors

Lithuanian-Russian Relations

May 1, 2024

Cyabra analyzed the narrative around Lithuanian-Russian relations on Facebook and X.

Diagram illustrating coordinated bot campaigns featuring colored network clusters of fake accounts, a sample fake profile with map, a propaganda post labelled NATO vs Russia, and a table listing suspicious profiles and networks

#National Security

Targets DEI Decision and The Inauthentic Backlash

May 5, 2025

Following Target's decision to scale back its DEI initiatives, Cyabra exposed a wave of inauthentic activity fueling outrage and economic threats.

Line graph of Target-related social media sentiment showing a 764% spike in late January 2025; metrics boxes list 2,262 engagements and 3,054,000 potential views, with two example tweets labeled fake beneath the chart

#Brand Reputation, Stocks, Threat Actors