#FactCheck-AI-Generated Image Falsely Shows Donald Trump Raising Fist During Washington Incident

Executive Summary 

‍

A dramatic video showing several people jumping from the upper floors of a building into what appears to be thick snow has been circulating on social media, with users claiming that it captures a real incident in Russia during heavy snowfall.  In the footage, individuals can be seen leaping one after another from a multi-storey structure onto a snow-covered surface below, eliciting reactions ranging from amusement to concern. The claim accompanying the video suggests that it depicts a reckless real-life episode in a snow-hit region of Russia.

A thorough analysis by CyberPeace confirmed that the video is not a real-world recording but an AI-generated creation. The footage exhibits multiple signs of synthetic media, including unnatural human movements, inconsistent physics, blurred or distorted edges, and a glossy, computer-rendered appearance. In some frames, a partial watermark from an AI video generation tool is visible. Further verification using the Hive Moderation AI-detection platform indicated that 98.7% of the video is AI-generated, confirming that the clip is entirely digitally created and does not depict any actual incident in Russia.

‍

Claim: 

‍

The video was shared on social media by an X (formerly Twitter) user ‘Report Minds’ on January 25, claiming it showed a real-life event in Russia. The post caption read: "People jumping off from a building during serious snow in Russia. This is funny, how they jumped from a storey building. Those kids shouldn't be trying this. It's dangerous." Here is the link to the post, and below is a screenshot. 

• https://x.com/reportminds/status/2015224546853343693?s=46 

‍

‍

Fact Check:

‍

The Desk used the InVid tool to extract keyframes from the viral video and conducted a reverse image search, which revealed multiple instances of the same video shared by other users with similar claims.  Upon close visual examination, several anomalies were observed, including unnatural human movements, blurred and distorted sections, a glossy, digitally-rendered appearance, and a partially concealed logo of the AI video generation tool ‘Sora AI’ visible in certain frames. Screenshots highlighting these inconsistencies were captured during the research .

• https://x.com/DailyLoud/status/2015107152772297086?s=20 
• https://x.com/75secondes/status/2015134928745164848?s=20 

‍

‍

The video was analyzed on Hive Moderation, an AI-detection platform, which confirmed that 98.7% of the content is AI-generated.

‍

‍

The viral video showing people jumping off a building into snow, claimed to depict a real incident in Russia, is entirely AI-generated. Social media users who shared it presented the digitally created footage as if it were real, making the claim false and misleading.

‍

Introduction

The increasing online interaction and popularity of social media platforms for netizens have made a breeding ground for misinformation generation and spread. Misinformation propagation has become easier and faster on online social media platforms, unlike traditional news media sources like newspapers or TV. The big data analytics and Artificial Intelligence (AI) systems have made it possible to gather, combine, analyse and indefinitely store massive volumes of data. The constant surveillance of digital platforms can help detect and promptly respond to false and misinformation content. 

During the recent Israel-Hamas conflict, there was a lot of misinformation spread on big platforms like X (formerly Twitter) and Telegram. Images and videos were falsely shared attributing to the ongoing conflict, and had spread widespread confusion and tension. While advanced technologies such as AI and big data analytics can help flag harmful content quickly, they must be carefully balanced against privacy concerns to ensure that surveillance practices do not infringe upon individual privacy rights. Ultimately, the challenge lies in creating a system that upholds both public security and personal privacy, fostering trust without compromising on either front.

The Need for Real-Time Misinformation Surveillance

According to a recent survey from the Pew Research Center, 54% of U.S. adults at least sometimes get news on social media. The top spots are taken by Facebook and YouTube respectively with Instagram trailing in as third and TikTok and X as fourth and fifth. Social media platforms provide users with instant connectivity allowing them to share information quickly with other users without requiring the permission of a gatekeeper such as an editor as in the case of traditional media channels. 

Keeping in mind the data dumps that generated misinformation due to the elections that took place in 2024 (more than 100 countries), the public health crisis of COVID-19, the conflicts in the West Bank and Gaza Strip and the sheer volume of information, both true and false, has been immense. Identifying accurate information amid real-time misinformation is challenging. The dilemma emerges as the traditional content moderation techniques may not be sufficient in curbing it. Traditional content moderation alone may be insufficient, hence the call for a dedicated, real-time misinformation surveillance system backed by AI and with certain human sight and also balancing the privacy of user's data, can be proven to be a good mechanism to counter misinformation on much larger platforms. The concerns regarding data privacy need to be prioritized before deploying such technologies on platforms with larger user bases. 

Ethical Concerns Surrounding Surveillance in Misinformation Control

Real-time misinformation surveillance could pose significant ethical risks and privacy risks. Monitoring communication patterns and metadata, or even inspecting private messages, can infringe upon user privacy and restrict their freedom of expression. Furthermore, defining misinformation remains a challenge; overly restrictive surveillance can unintentionally stifle legitimate dissent and alternate perspectives. Beyond these concerns, real-time surveillance mechanisms could be exploited for political, economic, or social objectives unrelated to misinformation control. Establishing clear ethical standards and limitations is essential to ensure that surveillance supports public safety without compromising individual rights. 

In light of these ethical challenges, developing a responsible framework for real-time surveillance is essential.

Balancing Ethics and Efficacy in Real-Time Surveillance: Key Policy Implications

Despite these ethical challenges, a reliable misinformation surveillance system is essential. Key considerations for creating ethical, real-time surveillance may include:

• Misinformation-detection algorithms should be designed with transparency and accountability in mind. Third-party audits and explainable AI can help ensure fairness, avoid biases, and foster trust in monitoring systems.
• Establishing clear, consistent definitions of misinformation is crucial for fair enforcement. These guidelines should carefully differentiate harmful misinformation from protected free speech to respect users’ rights.
• Only collecting necessary data and adopting a consent-based approach which protects user privacy and enhances transparency and trust. It further protects them from stifling dissent and profiling for targeted ads.
• An independent oversight body that can monitor surveillance activities while ensuring accountability and preventing misuse or overreach can be created. These measures, such as the ability to appeal to wrongful content flagging, can increase user confidence in the system.

Conclusion: Striking a Balance

Real-time misinformation surveillance has shown its usefulness in counteracting the rapid spread of false information online. But, it brings complex ethical challenges that cannot be overlooked such as balancing the need for public safety with the preservation of privacy and free expression is essential to maintaining a democratic digital landscape. The references from the EU’s Digital Services Act and Singapore’s POFMA underscore that, while regulation can enhance accountability and transparency, it also risks overreach if not carefully structured. Moving forward, a framework for misinformation monitoring must prioritise transparency, accountability, and user rights, ensuring that algorithms are fair, oversight is independent, and user data is protected. By embedding these safeguards, we can create a system that addresses the threat of misinformation and upholds the foundational values of an open, responsible, and ethical online ecosystem. Balancing ethics and privacy and policy-driven AI Solutions for Real-Time Misinformation Monitoring are the need of the hour.

References

• https://www.pewresearch.org/journalism/fact-sheet/social-media-and-news-fact-sheet/ 
• https://eur-lex.europa.eu/legal-content/EN/TXT/HTML/?uri=OJ:C:2018:233:FULL 

‍

Executive Summary

A recent viral message on social media such as X and Facebook, claims that the Indian Government will start charging an 18% GST on "good morning" texts from April 1, 2024. This news is misinformation. The message includes a newspaper clipping and a video that was actually part of a fake news report from 2018. The newspaper article from Navbharat Times, published on March 2, 2018, was clearly intended as a joke. In addition to this, we also found a video of ABP News, originally aired on March 20, 2018, was part of a fact-checking segment that debunked the rumor of a GST on greetings.

‍

Claims:

The claim circulating online suggests that the Government will start applying a 18% of GST on all "Good Morning" texts sent through mobile phones from 1st of April, this year. This tax would be added to the monthly mobile bills. 

‍

Fact Check:

When we received the news, we first did some relevant keyword searches regarding the news. We found a Facebook Video by ABP News titled Viral Sach: ‘Govt to impose 18% GST on sending good morning messages on WhatsApp?’

‍

We have watched the full video and found out that the News is 6 years old. The  Research Wing of CyberPeace Foundation also found the full version of the widely shared ABP News clip on its website, dated March 20, 2018. The video showed a newspaper clipping from Navbharat Times, published on March 2, 2018, which had a humorous article with the saying "Bura na mano, Holi hain." The recent viral image is a cutout image from ABP News that dates back to the year 2018. 

Hence, the recent image that is spreading widely is Fake and Misleading.

Conclusion:

The viral message claiming that the government will impose GST (Goods and Services Tax) on "Good morning" messages is completely fake. The newspaper clipping used in the message is from an old comic article published by Navbharat Times, while the clip and image from ABP News have been taken out of context to spread false information.

Claim: India will introduce a Goods and Services Tax (GST) of 18% on all "good morning" messages sent through mobile phones from April 1, 2024.

Claimed on: Facebook, X

Fact Check: Fake, made as Comic article by Navbharat Times on 2 March 2018