Fact Check: Viral Tiananmen Square Video Supporting Taiwan Independence Is AI-Generated

 

On August 23, 2026, a user named J4207Jin posted a 10-second video on X. The video appears to show a person holding a Taiwanese flag and chanting slogans in Tiananmen Square. The person can be heard saying, “Taiwan is not part of China, and it is independent.” The video has been viewed more than 30,000 times, reposted 13 times, and liked 64 times.

However, a subsequent investigation revealed that the video was not authentic and had been generated using artificial intelligence. To assess the authenticity of the footage, the video was examined using Gemini and HIVE Moderation.

The video was initially analyzed using Gemini, which determined that the content had been generated using artificial intelligence. Gemini identified several features as potential indicators of AI generation, including inconsistencies in lighting and shadows, unnatural movements of people in the video, and irregularities in the background.

The video was subsequently analyzed using HIVE Moderation, which also indicated that the footage was AI-generated.

 

Frames from the video were also examined using Google’s reverse image search to determine whether the footage had previously appeared online. The investigation revealed that the video was not first shared on August 23, 2026, but had circulated on different social media platforms several months earlier. The earliest instance identified during the investigation was an X post by a user named MinMin89485072 on January 1, 2026. The video was later shared on Instagram by a user named hellotaiwanese on January 4, 2026.

In conclusion, the investigation indicates that the viral video does not authentically depict a person holding a Taiwanese flag and making pro-Taiwan independence statements in Tiananmen Square. Analyses conducted using Gemini and HIVE Moderation identified indicators suggesting that the footage was generated using artificial intelligence. Furthermore, reverse image searches revealed that the video had circulated online months before the August 23, 2026 post, with the earliest identified version dating back to January 1, 2026. This case demonstrates how AI-generated content can be presented as footage of a real-world political event and potentially mislead viewers. It also highlights the importance of verifying both the authenticity and original context of viral videos, particularly when they involve politically sensitive issues.

 

 

If you suspect that a video, image, or audio file has been created using artificial intelligence or deepfake technology and would like free assistance in verifying its authenticity, you may send the link to the content or the file itself to allaboutdeepfake@gmail.com

AI-Generated Video Falsely Claims to Show Iranian Missiles Prepared for Attack on Israel

On August 24, 2026, a user named IRGC_Press posted a 15-second video on X. The video shows military vehicles transporting a large missile decorated with Iranian flags, accompanied by police escort vehicles.

The post accompanying the video stated: “Breaking: Iran announces it will use multi-warhead missiles with massive destructive power in waves 92 and 93 of its attacks on Israel tonight. It will be devastating for the Zionists, and other surprises have been prepared for them!”

 

The video and accompanying post have been viewed 5,382 times, reposted 55 times, and liked 159 times.

However, a subsequent investigation revealed that the video was not authentic and had been generated using artificial intelligence. IRGC_Press regularly shares false or misleading information using AI-generated content on X. To assess the authenticity of the footage, the video was examined using Grok and HIVE Moderation.

The video was initially analyzed using Grok, which determined that the content had been generated using artificial intelligence. It was subsequently analyzed using HIVE Moderation, which also indicated that the footage was AI-generated.

 

Frames from the video were also analyzed using Google’s reverse image search to determine whether the footage had previously appeared online. The analysis revealed that the video was first shared by a Facebook user on April 4, 2026, and was subsequently circulated on various other social media platforms at different times.

In conclusion, the investigation indicates that the video does not depict authentic footage of Iranian military vehicles transporting a multi-warhead missile for an imminent attack on Israel. Analyses conducted using Grok and HIVE Moderation suggest that the video was generated using artificial intelligence. Furthermore, reverse image searches revealed that the footage had circulated online months before the claim posted on August 24, 2026, further undermining the assertion that it showed preparations for an attack that night. This case demonstrates how AI-generated content can be combined with urgent and alarming claims to create misleading narratives and potentially influence public perceptions during periods of heightened geopolitical tension. It also highlights the importance of verifying both the authenticity and original context of viral visual content before accepting or sharing claims associated with it.

 

 

 

If you suspect that a video, image, or audio file has been created using artificial intelligence or deepfake technology and would like free assistance in verifying its authenticity, you may send the link to the content or the file itself to allaboutdeepfake@gmail.com

Real Video, False Claim: Disinformation in the Colombian Elections

The Colombian elections, held in two rounds—the first on May 31, 2026, and the second on June 21, 2026—were won by Abelardo de la Espriella, a right-wing candidate supported by U.S. President Donald Trump. Throughout the election process and its aftermath, many people posted on various social media platforms about the elections and their results. While some of these posts reflected factual information, others contained misleading content created for disinformation purposes.

One example of such misleading content is a video posted on X on June 22, 2026. The 13-second video was shared by SALVARANM, who claimed that supporters of candidate Iván Cepeda Castro burned Abelardo de la Espriella’s ballots in Plaza Mayor in Medellín. The video was viewed more than 250,400 times and received over 4,800 likes.

 

The video shared on X was first examined to determine its authenticity. To this end, it was analyzed using HIVE Moderation. The analysis revealed that the video contains real footage and was not generated by artificial intelligence. In other words, it was assessed as depicting a real event.

To better understand the context of the event and verify the accuracy of the claims, both Grok and Gemini were used for further analysis. Both tools concluded that while the video is authentic, the claims made in the post are false and misleading. Gemini specifically noted that the small burning object seen in the video is not an official ballot, pointing out that ballots used in Colombia are significantly larger.

Similarly, Grok found that although the video contains real footage, there is no evidence to support the claims made in the post. According to Grok, the assertion that the video shows ballots being burned cannot be verified, and the footage is being misrepresented.

In conclusion, the video contains real footage and does not appear to be AI-generated. However, investigations show that it does not depict the burning of election ballots belonging to Abelardo de la Espriella, as claimed. This example demonstrates that even authentic footage can be used to mislead the public when it is presented with false or unverified claims. It is therefore crucial to verify not only the authenticity of such footage but also the accuracy of the claims that accompany it.

Video Allegedly Showing Recent Attacks in Ukraine Is Actually From 2017

On June 21, 2026, a 19-second video was posted on X by @MesutUdaySaddam. The post claimed that the video showed Russia’s recent missile attacks on Ukraine on June 21 and the resulting explosions. The video has been viewed more than 81,300 times and has received 400 likes.

 

The video was investigated for authenticity because it contained elements that could suggest it was created using artificial intelligence. It was analyzed using Grok, Google SynthID, and HIVE Moderation tools. All three tools confirmed that the footage was real and not AI-generated. In other words, the video shared on X was determined to depict an actual event.

However, analyses by both Grok and Gemini indicate that the video does not show recent Russian attacks on Ukraine. Instead, it contains footage from an event that occurred in 2017 or earlier.

To determine the video’s actual time frame, it was further examined using Google’s reverse image search. This analysis revealed that the footage is from a fire and explosion that occurred on March 23, 2017, at a military ammunition depot in Balakliya, in Ukraine’s Kharkiv region.

In conclusion, the video is not AI-generated and contains real footage. However, it does not depict the Russian missile attacks on Ukraine on June 21, 2026, as claimed. Presenting footage from a different event in 2017 as if it were current demonstrates that the content is misleading. This example illustrates that even authentic footage can be used for disinformation when it is shared with incorrect information about the date, location, or event.

The Military Training Video Shared on TikTok Is Not Authentic

On March 31, 2026, a 10-second video was shared on TikTok by prizrak_kupyanska_zov. The post claimed that the footage depicted Russian soldiers undergoing rigorous military training. The video received 138 likes and was reposted 11 times.

 

However, subsequent investigation revealed that the video was not authentic and had been generated using artificial intelligence technologies. The authenticity of the footage was examined using BioID and HIVE Moderation.

The video was initially analyzed using HIVE Moderation, which determined that the content had been generated through artificial intelligence. It was subsequently examined using BioID, which reached the same conclusion, confirming that the footage was not genuine and had been synthetically produced.

 

In addition to the technical analyses, a detailed visual examination of the video revealed several inconsistencies commonly associated with AI-generated content. For example, while the soldiers shown during the first eight seconds of the video display Russian flags on their arms, the soldier appearing at the eighth second does not have this flag. This inconsistency within the same scene constitutes a significant indicator of artificial manipulation.

In conclusion, the video allegedly depicting Russian soldiers in training was generated using artificial intelligence technologies. Both the findings from technical detection tools and the observable visual inconsistencies demonstrate that the content does not represent authentic footage.

 

 

If you suspect that a video, image, or audio file has been created using artificial intelligence or deepfake technology and would like free assistance in verifying its authenticity, you may send the link to the content or the file itself to allaboutdeepfake@gmail.com.

The Video Allegedly Showing Captured Ukrainian Soldiers Was Generated Using Artificial Intelligence

On June 9, 2026, a 59-second video was posted on X by @AdrianoValui. The video purportedly shows Ukrainian soldiers taken prisoner by Russian forces. Within four hours of being shared, the footage had been viewed more than 28,000 times, received 573 likes, and was reposted 192 times.

 

 

However, subsequent investigation revealed that the video was not authentic and had been generated using artificial intelligence technologies. The authenticity of the footage was examined using Grok, the Media Analysis, Verification, and Retrieval Group (MeVer), and HIVE Moderation.

Initial analysis conducted with Grok suggested that the video was not genuine, indicating that it had likely been generated using AI and may have been circulated for pro-Russian propaganda purposes. Similarly, the assessment carried out by MeVer identified multiple visual inconsistencies commonly associated with AI-generated content, further supporting the conclusion that the video had been synthetically produced.

The footage was subsequently analyzed using HIVE Moderation, which confirmed that it had been generated using artificial intelligence, consistent with the findings from Grok and MeVer.

 

Moreover, the visible TikTok logo in the video indicates that the content was not originally posted on X and had previously circulated on other social media platforms. A reverse image search was therefore conducted, revealing that the video was first shared on TikTok on November 4, 2025, by “liliakv.” It was later posted on Facebook on December 16, 2025, by a user named “Petruss Krsska,” and continued to circulate through reposts across TikTok and Facebook on various dates.

 

 

In conclusion, the video allegedly depicting Ukrainian soldiers captured by Russian forces was generated using artificial intelligence technologies. The fact that the footage circulated across multiple platforms for several months demonstrates how AI-generated disinformation can persist over extended periods and be repeatedly presented as authentic. Particularly during times of war and conflict, it is essential to verify the source and accuracy of visual content using reliable sources and technical detection tools.

 

If you suspect that a video, image, or audio file has been created using artificial intelligence or deepfake technology and would like free assistance in verifying its authenticity, you may send the link to the content or the file itself to allaboutdeepfake@gmail.com.

 

Different Wars, Same Fake Video

On June 8, 2026, an 11-second video was posted on X by @IsraelSpoofX. The accompanying claim stated that Israeli air defense systems had detected and destroyed five Iranian warplanes. The video was viewed more than 1,600,000 times, received over 4,000 likes, and was reposted 926 times.

 

 

The same video was posted again on June 9, 2026, this time by @RussianArmy_ on X. However, the June 9 post claimed that the footage showed Russian air defense systems detecting and destroying five Ukrainian warplanes en route to attack Russia. Although both posts used identical footage, they attributed it to different conflicts and countries. The June 9 post received over 145,000 views, more than 1,400 likes, and 439 reposts.

However, the video shared by these two users and associated with two separate wars is not authentic; it was generated using artificial intelligence. The authenticity of the footage was examined using Grok, Google SynthID, and HIVE Moderation.

Initial analysis conducted with Grok indicated that both the video and the accompanying claims were not credible. Similarly, Google SynthID identified multiple visual and physical inconsistencies characteristic of AI-generated content. These findings strongly suggested that the footage had been synthetically produced.

 

The video was subsequently analyzed using HIVE Moderation, which confirmed that it was AI-generated, consistent with the results obtained from Grok and Google SynthID.

 

 

In conclusion, the video allegedly depicting Israel destroying Iranian warplanes and Russia destroying Ukrainian warplanes was generated using artificial intelligence. The recirculation of identical AI-generated content under different geopolitical narratives illustrates how easily disinformation can spread across social media platforms. This case once again underscores the importance of verifying visual content—particularly in relation to international conflicts and military developments—using reliable sources and technical detection tools.

 

 

If you suspect that a video, image, or audio file has been created using artificial intelligence or deepfake technology and would like free assistance in verifying its authenticity, you may send the link to the content or the file itself to allaboutdeepfake@gmail.com.

Videos of Soldiers Crying Posted on X Contain Disinformation

On June 8, 2026, a 15-second video was posted on X by @aattzz23. The video depicts a Ukrainian soldier, allegedly 23 years old, crying. The accompanying post claims that the soldier had been sent to the front lines and was therefore in distress. Within two hours, the video was viewed 732 times and received 34 likes.

 

 

However, subsequent investigation revealed that the video was not authentic and had been generated using artificial intelligence. The footage was analyzed using Google SynthID and HIVE Moderation, both AI-content detection tools. The results from both analyses confirmed that the video was AI-generated.

 

AI-generated videos portraying soldiers in emotional distress are not limited to the context of the war in Ukraine. Similar content has circulated on social media in relation to other conflicts. For example, on June 1, 2026, a nine-second video was shared on X by the user @DailyChinaNews. The video, which has been viewed more than 16,600 times and received 339 likes, shows a female U.S. soldier crying and stating that she misses her family.

 

Like the video shared on June 8, the June 1 video was also created using artificial intelligence. Analysis conducted using Google SynthID and HIVE Moderation determined that the footage was AI-generated. In particular, the SynthID analysis identified visual and physical inconsistencies characteristic of synthetic media.

 

In conclusion, both videos allegedly depicting Ukrainian and American soldiers crying were generated using artificial intelligence. These cases demonstrate how emotionally charged AI-generated content can be employed to influence public opinion during periods of war and conflict. It is therefore essential to verify the authenticity of content—especially material involving soldiers, civilians, or alleged victims of war—through reliable sources and technical detection tools before sharing it.

 

 

If you suspect that a video, image, or audio file has been created using artificial intelligence or deepfake technology and would like free assistance in verifying its authenticity, you may send the link to the content or the file itself to allaboutdeepfake@gmail.com.

The Image Allegedly Showing Protests in Tirana Was Generated Using Artificial Intelligence

Protests that began in Zvernec, Albania, on May 23, 2026, quickly spread to several cities across the country. On May 30, 2026, thousands of demonstrators gathered in the capital, Tirana, to protest the sale of a beach in Zvernec as part of a tourism project reportedly linked to U.S. President Donald Trump’s daughter, Ivanka Trump, and son-in-law, Jared Kushner. These protests have been publicly referred to as the “Flamingo Revolution.”

Content related to the protests in Albania has been widely circulated across various social media platforms. One such post was shared on June 4, 2026, by the user @SalisIlaria on X. The image posted by the user received over 52,000 views, more than 2,500 likes, and 734 reposts.

 

 

The image was claimed to depict tens of thousands of people participating in protests in Tirana. However, subsequent investigation revealed that the photograph was not authentic and had been generated using artificial intelligence. The image was analyzed using HIVE Moderation, which confirmed that it was AI-generated.

 

In conclusion, the photograph shared on X on June 4, 2026, allegedly showing protests in Tirana, was artificially generated. This case demonstrates how AI-generated images can be presented as authentic by associating them with ongoing political and social events. Therefore, it is essential to verify the accuracy of visual content—particularly in the context of politically sensitive and widely discussed events—by consulting reliable sources and using technical verification tools.

Images and Video Purporting to Show Iran Capturing Israeli and American Soldiers Were Generated Using Artificial Intelligence

On April 12, 2026, a 36-second video related to the war involving Iran, Israel, and the United States was posted on X under the username @MariaAlkaff_. The video purportedly shows Iranian soldiers capturing Israeli soldiers and depicts the treatment of the captured personnel. The post received 337 comments, was reposted 1,300 times, garnered 3,200 likes, and accumulated approximately 117,000 views.

 

However, the video is not authentic and was generated using artificial intelligence. It was initially analyzed using Google SynthID, which identified multiple indicators suggesting AI generation. In particular, inconsistencies in content, contextual coherence, and visual accuracy indicated that the footage had been produced using various AI-based tools.

In addition to SynthID, the video was examined using HIVE Moderation, another AI detection platform. The results from HIVE Moderation corroborated the initial findings, confirming that the video was generated through artificial intelligence.

 

 

A similar image was also shared on Facebook and Instagram on April 3, 2026. The accompanying claim alleged that the image depicted an F-35 pilot shot down by Iran on April 3, 2026, during interrogation.

Like the video shared on April 12, this image was also generated using artificial intelligence. Analysis conducted with SynthID revealed several anomalies—such as irregular physical features, lighting inconsistencies, and unusual environmental details within the interrogation setting—indicating AI production. Further examination using HIVE Moderation determined that the image was 92.5% likely to have been AI-generated.

 

In conclusion, both the image shared on April 3 and the video posted on April 12 are not authentic; both were created using artificial intelligence.

If you suspect that a video, image, or audio file has been created using artificial intelligence or deepfake technology and would like free assistance in verifying its authenticity, you may send the link to the content or the file itself to allaboutdeepfake@gmail.com.