#FactCheck -AI-Generated Clip Misleads Users as Monkey ‘Saves’ Child Under Falling Tree Branch
Executive Summary
A video showing a monkey allegedly saving the life of a sleeping child is rapidly going viral on social media. In the clip, a monkey can be seen picking up a child sleeping on a mat under a tree and moving the child away moments before a heavy tree branch falls at the same spot. Social media users are sharing the video as a “miracle of nature” and praising the emotional sensitivity and instincts of animals. However, research conducted by CyberPeace Research Wing found that the viral video is not real and was created using artificial intelligence tools.
Claim
The caption accompanying the viral post states:“In a shocking incident, a monkey was seen stepping in to save an innocent child sleeping under a tree from imminent danger. People nearby were stunned by the scene. It is being claimed that the monkey sensed the danger around the child and tried to protect him. The unusual incident has now gone viral on social media, with many saying that emotions and compassion are not limited to humans, animals can also understand feelings.”
The video has been widely shared across social media platforms
- https://www.instagram.com/reels/DYMvhRPTcCA/
- https://archive.ph/https://www.instagram.com/reels/DYMvhRPTcCA/
Fact Check
To verify the authenticity of the video, we extracted keyframes from the clip and conducted a reverse image search. During the research, we found the same video uploaded on May 8, 2026, on an Instagram page named Instagram user “mojilo_vandro.” The caption of the original upload did not provide any factual context and presented the video in a dramatic, miracle-like manner.
We further examined the Instagram account and found that it regularly posts several AI-generated videos featuring monkeys performing heroic or emotional acts. Importantly, the account owner has also identified themselves as an “AI video creator” in the bio section.
To further analyze the clip, we tested it using the AI detection tool Hive Moderation. The tool’s analysis classified the viral video as 85.6% likely to be AI-generated. We also checked the clip using another AI detection platform, Deepfake-o-meter. Its AVSRDD (2025) detection model flagged the video as potentially AI-generated with a 100% confidence score.
Conclusion
The evidence gathered during our research clearly shows that the viral video claiming to show a monkey saving a sleeping child from a falling tree branch is not authentic. The clip was created using AI-generated visual techniques and does not depict a real incident.
Related Blogs
Executive Summary
A video is being widely shared on social media showing what appears to be a missile attack over a densely populated city at night. The footage, seemingly recorded from inside a passenger aircraft, shows multiple explosions and missiles lighting up the sky.The video is being circulated with the claim that a Kuwaiti pilot filmed Iranian missiles being launched towards a US military base in the Middle East from the cockpit of an aircraft.
CyberPeace Research Wing’s research found the viral claim to be false. Our research confirmed that while Iran has launched attacks on US military installations in the Middle East, the viral footage is not related to those incidents. The video was found to be AI-generated and is being shared with a false and misleading narrative.
Claim
An Instagram user shared the viral video on July 11, 2026, claiming that it was recorded by a Kuwaiti pilot from an aircraft and shows Iranian missiles targeting a US military base in the Middle East.
Post Link: https://www.instagram.com/reel/DanymEDIdp2/?utm_source=ig_web_button_share_sheet
Fact Check
To verify the claim, we conducted a keyword-based search on Google. During the research, we did not find any credible news report or authentic source linking the viral video to the alleged Iranian missile attack. We did find a Navbharat Times report covering Iran's strikes on US military bases in the Middle East. However, the report did not contain the viral footage or any reference suggesting that such a video had been recorded by a Kuwaiti pilot. https://navbharattimes.indiatimes.com/video/news/america-iran-war-retaliation-for-us-attacks-begins-iran-strikes-at-trumps-allies-and-its-neighbour/videoshow/132370583.cms
In the next stage of the research, we analysed the viral video using the AI detection tool Hive Moderation. The tool indicated that the video had a 74.8% probability of being AI-generated.
For additional verification, we also analysed the footage using another AI detection tool, DetectVideo AI. Its analysis showed a 61% probability that the video was AI-generated.
Conclusion
Our research found that the viral claim is false. Although Iran has carried out missile strikes on US military facilities in the Middle East, the viral video is not authentic and has no connection to those attacks. The footage was found to be AI-generated and is being circulated on social media with a misleading claim that it was recorded by a Kuwaiti pilot from an aircraft.
Executive Summary:
Traditional Business Email Compromise(BEC) attacks have become smarter, using advanced technologies to enhance their capability. Another such technology which is on the rise is WormGPT, which is a generative AI tool that is being leveraged by the cybercriminals for the purpose of BEC. This research aims at discussing WormGPT and its features as well as the risks associated with the application of the WormGPT in criminal activities. The purpose is to give a general overview of how WormGPT is involved in BEC attacks and give some advice on how to prevent it.
Introduction
BEC(Business Email Compromise) in simple terms can be defined as a kind of cybercrime whereby the attackers target the business in an effort to defraud through the use of emails. Earlier on, BEC attacks were executed through simple email scams and phishing. However, in recent days due to the advancement of AI tools like WormGPT such malicious activities have become sophisticated and difficult to identify. This paper seeks to discuss WormGPT, a generative artificial intelligence, and how it is used in the BEC attacks to make the attacks more effective.
What is WormGPT?
Definition and Overview
WormGPT is a generative AI model designed to create human-like text. It is built on advanced machine learning algorithms, specifically leveraging large language models (LLMs). These models are trained on vast amounts of text data to generate coherent and contextually relevant content. WormGPT is notable for its ability to produce highly convincing and personalised email content, making it a potent tool in the hands of cybercriminals.
How WormGPT Works
1. Training Data: Here the WormGPT is trained with the arrays of data sets, like emails, articles, and other writing material. This extensive training enables it to understand and to mimic different writing styles and recognizable textual content.
2. Generative Capabilities: Upon training, WormGPT can then generate text based on specific prompts, as in the following examples in response to prompts. For example, if a cybercriminal comes up with a prompt concerning the company’s financial information, WormGPT is capable of releasing an appearance of a genuine email asking for more details.
3. Customization: WormGPT can be retrained any time with an industry or an organisation of interest in mind. This customization enables the attackers to make their emails resemble the business activities of the target thus enhancing the chances for an attack to succeed.
Enhanced Phishing Techniques
Traditional phishing emails are often identifiable by their generic and unconvincing content. WormGPT improves upon this by generating highly personalised and contextually accurate emails. This personalization makes it harder for recipients to identify malicious intent.
Automation of Email Crafting
Previously, creating convincing phishing emails required significant manual effort. WormGPT automates this process, allowing attackers to generate large volumes of realistic emails quickly. This automation increases the scale and frequency of BEC attacks.
Exploitation of Contextual Information
WormGPT can be fed with contextual information about the target, such as recent company news or employee details. This capability enables the generation of emails that appear highly relevant and urgent, further deceiving recipients into taking harmful actions.
Implications for Cybersecurity
Challenges in Detection
The use of WormGPT complicates the detection of BEC attacks. Traditional email security solutions may struggle to identify malicious emails generated by advanced AI, as they can closely mimic legitimate correspondence. This necessitates the development of more sophisticated detection mechanisms.
Need for Enhanced Training
Organisations must invest in training their employees to recognize signs of BEC attacks. Awareness programs should emphasise the importance of verifying email requests for sensitive information, especially when such requests come from unfamiliar or unexpected sources.
Implementation of Robust Security Measures
- Multi-Factor Authentication (MFA): MFA can add an additional layer of security, making it harder for attackers to gain unauthorised access even if they successfully deceive an employee.
- Email Filtering Solutions: Advanced email filtering solutions that use AI and machine learning to detect anomalies and suspicious patterns can help identify and block malicious emails.
- Regular Security Audits: Conducting regular security audits can help identify vulnerabilities and ensure that security measures are up to date.
Case Studies
Case Study 1: Financial Institution
A financial institution fell victim to a BEC attack orchestrated using WormGPT. The attacker used the tool to craft a convincing email that appeared to come from the institution’s CEO, requesting a large wire transfer. The email’s convincing nature led to the transfer of funds before the scam was discovered.
Case Study 2: Manufacturing Company
In another instance, a manufacturing company was targeted by a BEC attack using WormGPT. The attacker generated emails that appeared to come from a key supplier, requesting sensitive business information. The attack exploited the company’s lack of awareness about BEC threats, resulting in a significant data breach.
Recommendations for Mitigation
- Strengthen Email Security Protocols: Implement advanced email security solutions that incorporate AI-driven threat detection.
- Promote Cyber Hygiene: Educate employees on recognizing phishing attempts and practising safe email habits.
- Invest in AI for Defense: Explore the use of AI and machine learning in developing defences against generative AI-driven attacks.
- Implement Verification Procedures: Establish procedures for verifying the authenticity of sensitive requests, especially those received via email.
Conclusion
WormGPT is a new tool in the arsenal of cybercriminals which improved their options to perform Business Email Compromise attacks more effectively and effectively. Therefore, it is critical to provide the defence community with information regarding the potential of WormGPT and its implications for enhancing the threat landscape and strengthening the protection systems against advanced and constantly evolving threats.
This means the development of rigorous security protocols, general awareness of security solutions, and incorporating technologies such as artificial intelligence to mitigate the risk factors that arise from generative AI tools to the best extent possible.
Introduction
In a significant step, the Indian Army beefed up its information warfare capacity on June 25, 2026, with the operationalisation of @MythbusterXX, its dedicated fact-checking handle designed to counter any form of misinformation, disinformation, malinformation, and deepfakes on the army swiftly. Adopting the motto 'Verify Before You Amplify,' the service seeks to pivot from reactive statements to active cognitive warfare. In a milieu where manipulated narratives can be as decisive in shaping public perception as kinetic force is on the battlefield, truth itself has transformed into a critical national security objective.
This is clear proof that protecting India’s digital battleground will from now on be defined not just by troop deployments but also by its institutional verification capacities, swift attribution mechanisms, and public awareness.
When the Battlefield Went Digital
Today warfare extends to timelines, group chats, and prime-time graphics packages. The distinction between misinformation , disinformation , and malinformation is crucial to operations. After all, they necessitate different types of counters. Generative AI just exponentially increased their velocity, cost, and the creepy believability of synthesized audio, video, and images. Operation Sindoor, India's May 2025 military response to a Pahalgam terror attack, provides a blueprint for just how large it can get.
According to the fact-checking site BOOM, 68% of fact checks in May related to Operation Sindoor, describing the campaign as a misinformation superspreader, and, more directly, India's Chief of Defence Staff General Anil Chauhan lamented at last year’s Shangri-La Dialogue that roughly 15% of his military’s operational time was spent on countering false news. New Delhi matched this response level; the Ministry of Information and Broadcasting had blocked over 1,400 URLs, many bearing false information or communally inciteful narratives from accounts in Pakistan. These challenges won't be vanishing any time soon. Microsoft's July 2025 Digital Defence Report lists India among the top countries targeted for AI-powered state-backed hacking, noting the automation of attacks and generation of fake content used to influence opinion.
CyberPeace's regulatory tracking also indicates steps toward building infrastructure to mitigate this, a Rule 7 complaint system for deepfakes, and efforts to foster local detection capabilities under the IndiaAI mission signal it's now an infrastructural threat.
Why the Army Chose to Speak First
On 1 June 2026, the Indian government’s official fact-checking unit debunked a deepfake of former Army Chief General Dhiraj Seth talking about India’s engagement with the Taliban that emanated from Pakistan-affiliated propaganda sources, appearing within days of a change of military command. Barely days after General Dhiraj Seth was elevated to Chief of Army Staff, the 31st person to hold the post, another fabricated deepfake used spliced authentic footage with AI-cloned audio to falsely allege he had blamed the past army leadership for hiding the bodies of soldiers to protect their image.
Such a dual targeting of India’s army chiefs with deepfakes in such close succession is a testament as to why an authoritative, constantly running fact-check machine isn’t an optional extra but a strategic must-have.
Architecture is as important as architecture itself. India tried a statutory fact-checking model with a government-owned Fact Check Unit under the Information Technology (IT) Rules of 2023 through the Press Information Bureau (PIB). It had to retreat from the edge of the constitutional cliff. In September 2024, when the Bombay High Court struck down key provisions of these rules. It ruled it unconstitutional to empower the state to declare digital content relating to the state itself as fake, false, or misleading. A similar reading of the analysis by CyberPeace found that digital speech should be given the same protections as offline speech, and this should be the benchmark for any counter-misinformation policy.
MythbusterXX carefully avoids the constitutional controversy by acting as an institutional avenue that produces verified information and rebuttals in response to disinformation. There is a profound distinction between answering falsehoods with factual, credible speech versus shutting them down with state power, the very proportionality that civil libertarians have said must govern a democracy's approach to online misinformation.
Platforms as Accelerants, Not Just Conduits
Part of the issue lies upstream, with the algorithms and broadcasters that favor speedy falsehood over careful confirmation. A month’s supply of “false and misleading stories,” the Reuters Institute for the Study of Journalism observed, streamed online within hours of Operation Sindoor's onset: one fact-checker identified approximately seventy false claims by the close of day one. Meanwhile, on the small screen, graphics depicting escalating conflicts were being televised, with broadcasters forced to scramble under 24/7 competition.
Amplified manipulation of the crisis narrative is also a factor; studies show that thousands of accounts have been reposting exactly the same party content for three years, making crisis-induced viral narratives predictable rather than organic phenomena. For the most part, the platforms that manage these narratives and the algorithms that direct content on both sides of the divide are not just innocent delivery systems.
Digital Literacy as the First Line of Defence
Institutional rebuttals only go so far if citizens lack the reflex to pause before sharing, which is where CyberPeace's own work becomes directly relevant. Through a multi-year, Google.org-backed initiative, CyberPeace Foundation aims to reach over 40 million Indian internet users, including 9 million underserved beneficiaries, through a multilingual resource centre offering 650 hours of content, state-level helplines, and quick-response teams staffed by digital forensics and fact-checking experts. Longer-running efforts such as the Digital Shakti campaign and the annual eRaksha competition, run with NCERT, have built a culture of responsible digital citizenship among young and first-time internet users since 2019, while CyberPeace Corps, the foundation's volunteer arm, carries the same message into classrooms and campuses through cyber-awareness sessions run with universities and school networks nationwide. Notably, CyberPeace's own research on children's online safety had already flagged manipulated Army-related videos as a category of digital manipulation designed to cast doubt on official military positions, well before @MythbusterXX existed. The Army's tagline and CyberPeace's mission converge on the same insight: verification is a civic skill, not merely an institutional service.
A Season of Deepfakes
The history book of compromised defence material in the past year alone would be edifying. The International Federation of Journalists recorded one such deepfake widely circulated that relied on AI voice cloning and lip-sync tools to present the Pakistani PM conceding defeat, though the video actually contained his praise for the Pakistani air force’s performance following Operation Sindoor.
Other similar videos featured the Indian Prime Minister, External Affairs Minister, and Home Minister allegedly apologising to Pakistan and a deepfake with a foreign head of state in voice-cloned style applauding India’s armed forces. None needed a state’s resources, but just a laptop, voice cloning available readily, and a citizenry ready to spread rather than confirm the authenticity of any such sensational information before spreading it on to the masses.
Building the Verification Reflex
A resilient information ecosystem needs different actors playing complementary roles:
- Citizens: Do not equate virality with credibility. Verify all national security-related information through official sources, such as @MythbusterXX, before reposting or forwarding it.
- Journalists: Apply rigorous verification standards to live broadcasts, war-room graphics, and breaking reports, just as you would to print journalism, and avoid speculation during fast-moving military operations.
- Researchers and fact-checkers: Use open-source forensic and AI-detection technologies to authenticate questionable material. BOOM researchers, for instance, used Deepfake-o-meter to scrutinize misleading videos of political leaders before fact-checking and publishing them.
- Policymakers: Favored constitutionally proportional and carefully tailored regulations over broad-based takedown mandates. While the 2026 IT Amendment Rules require platforms to shift their obligations from compliance with takedowns towards preventive diligence on synthetic media, implementation should continue to be informed by judicial protections afforded to freedom of speech and due process.
Cyber Resilience Is National Security Now
Legal scholars examining Operation Sindoor have drawn a useful distinction between coordinated information warfare, which is strategic and intentional, and the diffuse, uncoordinated mis/disinformation that dominated timelines during the conflict, cautioning that disproportionate state responses to the latter can compromise citizens' right to know just as much as the falsehoods themselves. Getting that balance right, between speed and due process, between institutional voice and censorship, is the real test facing India's information ecosystem, in defence and far beyond it. CyberPeace has made a related argument in its own work on AI-enabled espionage: institutions such as the National Critical Information Infrastructure Protection Centre and the Defence Cyber Agency are already folding AI-based monitoring into their processes, yet no amount of institutional surveillance substitutes for a citizenry trained to spot manipulation on sight.
@MythbusterXX will not end deepfakes, and no single handle can. But it signals something CyberPeace has argued for years: resilience against synthetic and manipulated media requires authoritative institutional voices, digitally literate citizens, forensically equipped researchers, and proportionate policy, all pulling in the same direction. "Verify before you amplify" is not just an Army campaign. It is the operating discipline a democracy needs to protect its own information age.
Conclusion
@MythbusterXX can't possibly kill deepfakes and disinformation, but this marks a milestone shift towards developing an institutional resilience for India's info battlefield. For, after all, national security in the age of info warfare depends less on tech and more on robust institutions, constitutional balance in policymaking, responsible social platforms, and a citizenry that clicks "forward" only after clicking "verify."
* * * * *
This piece is part of CyberPeace's ongoing work on misinformation, disinformation, and digital citizenship in India. For more on CyberPeace's initiatives in digital literacy and cyber resilience, visit cyberpeace.org.
Key Sources
- Dynamite News, "Indian Army launches fact-check push against deepfake videos and fake military claims" (June 2026)
- ADG PI – Indian Army, official announcement on X (@adgpi)
- Press Information Bureau, Government of India, press release on countering misinformation during Operation Sindoor
- Reuters Institute for the Study of Journalism, "Truth is the casualty: How Indian fact-checkers debunked false claims during the India-Pakistan crisis"
- International Federation of Journalists, "AI, Deepfakes, and the Fog of War" and "Operation Sindoor and the Two Wars" (June 2025)
- Republic World, "Operation Sindoor Haunts Pakistan: Islamabad's Latest AI Deepfake Bid Against Indian Army Chief General Dhiraj Seth Exposed" (July 2026)
- TechPolicy Press, "Sanity Prevails as Bombay High Court Strikes Down India Government's Fact Check Unit", and LiveLaw, coverage of the tie-breaker verdict, on the IT Amendment Rules, 2023 (Fact-Check Unit)
- Dark Reading, "Indian Army Propaganda Spread by 1.4K AI-Powered Social Media Accounts"
- Inforrm, "(Dis)information warfare and the right to know: lessons from Operation Sindoor"
- CyberPeace Foundation, initiatives page and The CyberPeace Initiative (Google.org-backed digital literacy programme)
- CyberPeace Foundation, "AI-Powered Espionage: How India's Cybersecurity Strategy Must Evolve"
- CyberPeace Foundation, "From Deepfakes to Due Diligence – Decoding India's IT Amendment Rules"
- Wikipedia, CyberPeace Foundation (background on Digital Shakti and eRaksha)
