#FactCheck: Fake viral AI video captures a real-time bridge failure incident in Bihar

Executive Summary
‍

A video is being shared on social media showing a group of people dancing on a road while carrying saffron flags. A mosque can also be seen nearby in the video Sharing this clip, some users are claiming that it is from Uttam Nagar in Delhi, where members of the Hindu community celebrated Holi on the occasion of Eid on March 21. Research by the CyberPeace found the viral claim to be misleading. Our probe revealed that the video is not related to Holi celebrations on Eid in Uttam Nagar, Delhi. In fact, the video has been available on the internet since 2024 and is said to be from Raichur district in Karnataka. Several users have shared it claiming that it was recorded during Ganesh Chaturthi celebrations.

‍

Claim:
‍

A social media user shared the viral video on March 21, 2026, with a misleading claim. The link and archive link of the post are given below.

• https://x.com/biharigurl/status/2035651541420687711 
• https://archive.ph/5cxgW 

‍

‍

Fact Check:
‍

To verify the viral claim, we first conducted a keyword search on Google. However, we did not find any credible media report supporting the claim. In the next step, we extracted keyframes from the video and performed a reverse search using Google Lens. During this process, we found the same video on an Instagram account, which was posted on September 23, 2024.

• https://www.instagram.com/p/DAQ81GSqY3u/ 

‍

The user had captioned the video as “Ganesh Chaturthi 2024,” suggesting that the clip is related to the festival. Further, upon closely analyzing the video, we noticed that the mosque visible in the background had “Usmania Masjid” written on it. We then searched for this location on Google Maps and found that the mosque is located on Teen Khandil Road in Raichur, Karnataka, which matches the visuals seen in the viral clip.

‍

‍

Conclusion:
‍

Our research found that the video is not from Uttam Nagar, Delhi, nor is it related to Holi celebrations on Eid. The clip has been available online since 2024 and is from Raichur, Karnataka. It has been shared with a misleading claim and is actually linked to Ganesh Chaturthi celebrations.

‍

Introduction

In 2025, the internet is entering a new paradigm and it is hard not to witness it. The internet as we know it is rapidly changing into a treasure trove of hyper-optimised material over which vast bot armies battle to the death, thanks to the amazing advancements in artificial intelligence. All of that advancement, however, has a price, primarily in human lives. It turns out that releasing highly personalised chatbots on a populace that is already struggling with economic stagnation, terminal loneliness, and the ongoing destruction of our planet isn’t exactly a formula for improved mental health. This is the truth of 75% of the kids and teen population who have had chats with chatbot-generated fictitious characters. AI, or artificial intelligence, Chatbots are becoming more and more integrated into our daily lives, assisting us with customer service, entertainment, healthcare, and education. But as the impact of these instruments grows, accountability and moral behaviour become more important. An investigation of the internal policies of a major international tech firm last year exposed alarming gaps: AI chatbots were allowed to create content with child romantic roleplaying, racially discriminatory reasoning, and spurious medical claims. Although the firm has since amended aspects of these rules, the exposé underscores an underlying global dilemma - how can we regulate AI to maintain child safety, guard against misinformation, and adhere to ethical considerations without suppressing innovation? 

The Guidelines and Their Gaps

The tech giants like Meta and Google are often reprimanded for overlooking Child Safety and the overall increase in Mental health issues in children and adolescents. According to reports, Google introduced Gemini AI kids, a kid-friendly version of its Gemini AI chatbot, which represents a major advancement in the incorporation of generative artificial intelligence (Gen-AI) into early schooling. Users under the age of thirteen can use supervised accounts on the Family Link app to access this version of Gemini AI Kids.

AI operates on the premise of data collection and analysis. To safeguard children’s personal information in the digital world, the Digital Personal Data Protection Act, 2023 (DPDP Act) introduces particular safeguards. According to Section 9, before processing the data of children, who are defined as people under the age of 18, Data Fiduciaries, entities that decide the goals and methods of processing personal data, must get verified consent from a parent or legal guardian. Furthermore, the Act expressly forbids processing activities that could endanger a child’s welfare, such as behavioural surveillance and child-targeted advertising. According to court interpretations, a child's well-being includes not just medical care but also their moral, ethical, and emotional growth. 

While the DPDP Act is a big start in the right direction, there are still important lacunae in how it addresses AI and Child Safety. Age-gating systems, thorough risk rating, and limitations specific to AI-driven platforms are absent from the Act, which largely concentrates on consent and damage prevention in data protection. Furthermore, it ignores the threats to children’s emotional safety or the long-term psychological effects of interacting with generative AI models. Current safeguards are self-regulatory in nature and dispersed across several laws, such as the Bhartiya Nyaya Sanhita, 2023. These include platform disclaimers, technology-based detection of child-sexual abuse content, and measures under the IT (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021. 

Child Safety and AI

• The Risks of Romantic Roleplay - Enabling chatbots to engage in romantic roleplaying with youngsters is among the most concerning discoveries. These interactions can result in grooming, psychological trauma, and relaxation to inappropriate behaviour, even if they are not explicitly sexual. Having illicit or sexual conversations with kids in cyberspace is unacceptable, according to child protection experts. However, permitting even "flirtatious" conversation could normalise risky boundaries.
•  International Standards and Best Practices - The concept of "safety by design" is highly valued in child online safety guidelines from around the world, including UNICEF's Child Online Protection Guidelines and the UK's Online Safety Bill. This mandating of platforms and developers to proactively remove risks, not reactively to respond to harms, is the bare minimum standard that any AI guidelines must meet if they provide loopholes for child-directed roleplay.

Misinformation and Racism in AI Outputs

• The Disinformation Dilemma - The regulations also allowed AI to create fictional narratives with disclaimers. For example, chatbots were able to write articles promulgating false health claims or smears against public officials, as long as they were labelled as "untrue." While disclaimers might give thin legal cover, they add to the proliferation of misleading information. Indeed, misinformation tends to spread extensively because users disregard caveat labels in favour of provocative assertions.

•   Ethical Lines and Discriminatory Content - It is ethically questionable to allow AI systems to generate racist arguments, even when requested. Though scholarly research into prejudice and bias may necessitate such examples, unregulated generation has the potential to normalise damaging stereotypes. Researchers warn that such practice brings platforms from being passive hosts of offensive speech to active generators of discriminatory content. It is a difference that makes a difference, as it places responsibility squarely on developers and corporations.

The Broader Governance Challenge

• Corporate Responsibility and AI Material generated by AI is not equivalent to user speech—it is a direct reflection of corporate training, policy decisions, and system engineering. This fact requires a greater level of accountability. Although companies can update guidelines following public criticism, that there were such allowances in the first place indicates a lack of strong ethical regulation.
•  Regulatory Gaps Regulatory regimes for AI are currently in disarray. The EU AI Act, the OECD AI Principles, and national policies all emphasise human rights, transparency, and accountability. The few, though, specify clear guidelines for content risks such as child roleplay or hate narratives. This absence of harmonised international rules leaves companies acting in the shadows, establishing their own limits until contradicted.

An active way forward would include

• Express Child Protection Requirements: AI systems must categorically prohibit interactions with children involving flirting or romance.
• Misinformation Protections: Generative AI must not be allowed to generate knowingly false material, disclaimers being irrelevant.
• Bias Reduction: Developers need to proactively train systems against generating discriminatory accounts, not merely tag them as optional outputs.
• Independent Regulation: External audit and ethics review boards can supply transparency and accountability independent of internal company regulations.

Conclusion

The guidelines that are often contentious are more than the internal folly of just one firm; they point to a deeper systemic issue in AI regulation. The stakes rise as generative AI becomes more and more integrated into politics, healthcare, education, and social interaction. Racism, false information, and inadequate child safety measures are severe issues that require quick resolution. Corporate regulation is only one aspect of the future; other elements include multi-stakeholder participation, stronger global systems, and ethical standards. In the end, rather than just corporate interests, trust in artificial neural networks will be based on their ability to preserve the truth, protect the weak, and represent universal human values.

References

• https://www.esafety.gov.au/newsroom/blogs/ai-chatbots-and-companions-risks-to-children-and-young-people 
•   https://www.lakshmisri.com/insights/articles/ai-for-children/# 
•   https://the420.in/meta-ai-chatbot-guidelines-child-safety-racism-misinformation/
• https://www.unicef.org/documents/guidelines-industry-online-child-protection
•  https://www.oecd.org/en/topics/sub-issues/ai-principles.html
• https://artificialintelligenceact.eu/

‍

Executive Summary:

A viral photo on social media claims to show a ruined bridge in Kerala, India. But, a reality check shows that the bridge is in Amtali, Barguna district, Bangladesh. The reverse image search of this picture led to a Bengali news article detailing the bridge's critical condition. This bridge was built-in 2002 to 2006 over Jugia Khal in Arpangashia Union. It has not been repaired and experiences recurrent accidents and has the potential to collapse, which would disrupt local connectivity. Thus, the social media claims are false and misleading.

Claims:

Social Media users share a photo that shows a ruined bridge in Kerala, India. 

Fact Check:

On receiving the posts, we reverse searched the image which leads to a Bengali News website named Manavjamin where the title displays, “19 dangerous bridges in Amtali, lakhs of people in fear”. We found the picture on this website similar to the viral image. On reading the whole article, we found that the bridge is located in Bangladesh's Amtali sub-district of Barguna district.

Taking a cue from this, we then searched for the bridge in that region. We found a similar bridge at the same location in Amtali, Bangladesh.

According to the article, The 40-meter bridge over Jugia Khal in Arpangashia Union, Amtali, was built in 2002 to 2006 and was never repaired. It is in a critical condition, causing frequent accidents and risking collapse. If the bridge collapses it will disrupt communication between multiple villages and the upazila town. Residents have made temporary repairs. 

Hence, the claims made by social media users are fake and misleading.

Conclusion:

In conclusion, the viral photo claiming to show a ruined bridge in Kerala is actually from Amtali, Barguna district, Bangladesh. The bridge is in a critical state, with frequent accidents and the risk of collapse threatening local connectivity. Therefore, the claims made by social media users are false and misleading.

• Claim: A viral image shows a ruined bridge in Kerala, India. 
• Claimed on: Facebook
• Fact Check: Fake & Misleading

‍