#FactCheck-Mock drill video falsely linked to death of Indian sailor in Gulf of Oman
Executive Summary
Amid reports of attacks on ships in the Gulf of Oman that led to the death of three Indian nationals, a video showing a person lying injured on a ship is being widely circulated on social media. The clip is being shared with the claim that the United States sent the body of a deceased Indian sailor to India in a “wrapped” condition. CyberPeace Research Wing research found the claim to be misleading. The viral video is unrelated to any recent incident involving Indian nationals and is, in fact, from a mock drill conducted on board a ship.
Claim:
An Instagram user ‘fakirchand.sharma’ shared the video on June 15, 2026, claiming that an Indian sailor, identified as Nishant Urthnathan, died due to health complications on a vessel anchored at Duqm port, Oman. The post further alleged that the United States returned the body in a “wrapped” condition, sparking outrage on social media. https://www.instagram.com/fakirchand.sharma/reel/DZlsxj9TIaS , https://perma.cc/C9SF-J8ZR
Fact Check:
A reverse image search of keyframes from the viral video led to an Instagram account ‘bipul_raja_vlogs’, where the same video was posted on June 10, 2026. The original caption clearly stated that the visuals were from a fire drill conducted on board a ship. https://www.instagram.com/p/DZbnZlGPAlL
Further posts from the same account on June 15 and June 17, 2026, also show similar mock drill visuals and clarify that the content is being misused online. The account features multiple posts related to routine ship operations and training exercises. https://www.instagram.com/p/DZnEXOzCc2X
The profile also contains several other photos and videos of crew members working on the ship. We also checked the YouTube channel linked in this profile, “Sunnybabu,” where a mock drill video was uploaded on January 21, 2026. https://www.instagram.com/bipul_raja_vlogs
Conclusion:
The research confirms that the viral video is unrelated to the death of any Indian sailor in the Gulf of Oman. The clip actually shows a mock drill conducted on a ship and is being falsely shared with a misleading narrative.
Related Blogs
Executive Summary
A news graphic is being widely shared on social media claiming that Union Education Minister Dharmendra Pradhan will resign on July 22. The graphic quotes him as saying, "Respecting the sentiments of the country's youth, I have taken this decision." CyberPeace Research Wing ’s research found the claim to be false. The probe revealed that no official announcement has been made regarding Dharmendra Pradhan’s resignation as Union Education Minister. However, demands for his resignation have intensified from opposition parties and student groups over alleged irregularities in NEET and other examinations, leading to protests in Delhi and other places. The research also found that the viral news graphic circulating on social media was likely created using Artificial Intelligence (AI).
Claim:
A social media user on Instagram shared the viral news graphic on July 21, 2026, claiming that Union Education Minister Dharmendra Pradhan is set to resign on July 22. The post also claimed that Pradhan said during a press conference, "Respecting the sentiments of the country's youth, I have taken this decision."
https://www.instagram.com/reel/DbEMoRtgLfT/?igsh=c3g3dzZxZ2U3YW01
Fact Check:
To verify the authenticity of the viral claim, we conducted a Google search using relevant keywords. However, we did not find any credible media reports confirming that Union Education Minister Dharmendra Pradhan is resigning on July 22.Further, we examined the official X (formerly Twitter) account of Union Education Minister Dharmendra Pradhan. No official announcement or post related to his resignation was found on his account.
https://x.com/dpradhanbjp?lang=en
Upon examining the viral news graphic, we noticed several indicators suggesting that it could be AI-generated. To verify this, we analysed the graphic using the AI detection tool AI or Not. The tool’s analysis indicated a 94% probability that the graphic was generated using AI.
We also scanned the viral graphic through another AI detection tool, WasIt AI. According to the tool’s results, the probability of the graphic being AI-generated was found to be 87%.
Conclusion:
CyberPeace Research Wing ’s fact check found the viral claim to be false. No official announcement has been made regarding Union Education Minister Dharmendra Pradhan’s resignation, and reports claiming that he will resign on July 22 are incorrect. The research further revealed that the viral news graphic circulating on social media was created using Artificial Intelligence (AI).
Introduction
Agentic AI systems are autonomous systems that can plan, make decisions, and take actions by interacting with external tools and environments. But they shift the nature of risk by blurring the lines among input, decision, and execution. A conventional model generates an output and stops. An agent takes input, makes plans, invokes tools, updates its state and repeats the cycle. This creates a system where decisions are continuously revised through interaction with external tools and environments, rather than being fixed at the point of input.
This means the attack surface expands in size and becomes more dynamic. Instead of remaining confined to components as in traditional computational systems, they spread in layers and can continue to grow through time. To understand this shift, the system can be analysed through functional layers such as inputs, memory, reasoning, and execution, while recognising that risk does not remain isolated within these layers but emerges through their interaction.
Agentic AI Attack Surface
A layered view of how risks emerge across input, memory, reasoning, execution, and system integration, including feedback loops and cross-system dependencies that amplify vulnerabilities.
Input Layer: Where Untrusted Data Becomes Control
The entry point of an agent is no longer one prompt. The documents, APIs, files, system logs and the outputs of other agents can now be considered input. This diversity is significant due to the fact that every source of input carries its own trust assumptions, and in the majority of cases, they are weak.
The most obvious threat is prompt injection, where inputs are treated as instructions rather than data. Since inputs are treated as instructions, a virus, a malicious webpage, or a document can contain instructions that override system goals without necessarily being detected as something harmful.
Indirect prompt injection extends this risk beyond direct user interaction. Instead of targeting the interface, attackers compromise the retrieval process by embedding malicious instructions within external data sources. When the agent retrieves and processes the data, it treats the embedded content as legitimate input. As a result, the attack is executed through normal reasoning processes, allowing the system to act on untrusted data without recognising the manipulation.
Data poisoning also occurs at runtime. In contrast to classical poisoning (where training data is manipulated), runtime poisoning distorts the agent’s perception of its environment as it runs. This can change decisions without causing apparent failures.
Obfuscation introduces another indirect attacker vector. Encoded instructions or complicated forms may bypass human review but remain readable to the model. This creates asymmetry whereby the system knows more about the attack than those operating it. Once compromised at this layer, the agent implements compromised instructions which affect downstream operations.
Context and Memory: Persistence of Influence
Agentic systems depend on memory to operate efficiently. They often retain context across sessions and frequently store information between sessions.
This introduces a different type of risk: persistence. Through memory poisoning, attackers can insert false or adversarial information into sorted context, which then influences future decisions. Unlike prompt injection, which is often limited to a single interaction, this effect carries forward. Over time, the agent begins to operate on a distorted internal state, shaping decisions in ways that may not be immediately visible.
Another issue is cross-session leakage. Information in a particular context may be replayed in a different context when memory is being shared or there is insufficient memory separation. This is specifically dangerous in those systems that combine retrieval and long-term storage. The context management in itself becomes a weakness. Agents are required to make decisions on what to retain and what to discard. This is susceptible to attackers who can flood the context or manipulate what is still visible and indirectly affect reasoning.
The underlying problem is structural. Memory turns data into a state. Once state is corrupted, the system cannot easily distinguish valid knowledge from adversarial influence.
The issue is structural. Memory converts temporary data into a persistent state. Once this state is weakened, the system cannot reliably separate valid information from adversarial influence, making recovery significantly more difficult.
Reasoning and Planning: Manipulating Intent Without Breaking Logic
The reasoning layer is where agentic AI stands apart from traditional systems. The model no longer reacts to inputs alone. It actively breaks down objectives, analyses alternatives, and ranks actions.
At the reasoning stage, the nature of risk shifts. The concern is no longer limited to injecting instructions, but to influencing how decisions are made. One example is goal manipulation, where the agent subtly reinterprets its objective and produces outcomes that are technically correct but strategically harmful. Reasoning hijacking operates within intermediate steps, altering how constraints are evaluated or how trade-offs are prioritised. The system may remain internally consistent, which makes such deviations difficult to detect.
Tool selection becomes a critical control point. Agents decide which tools to use and when, so influencing these choices can redirect execution without directly accessing the tools themselves. Hallucinations also take on a different role here. In static systems, they remain errors. In agentic systems, they can trigger actions. A perceived need or incorrect judgement can translate into real-world consequences.
This layer introduces probabilistic failure. The system is not fully weakened, but it is nudged towards decisions that appear reasonable yet are incorrect. The risk lies in how those decisions are justified.
Tool and Execution: When Decisions Gain Reach
Once an agent begins interacting with tools, its behaviour extends beyond the model into external systems. APIs, databases, and services become part of the execution path.
One key risk is the use of unauthorised tools. When agents operate with broad permissions, any manipulation of the upstream can be converted into real-world actions. This makes access control a central security concern. Command injection also takes a different form here. The agent generates commands based on its reasoning, so if that reasoning is compromised, the resulting actions may still appear valid despite being harmful.
External tool outputs introduce another risk. If these systems return corrupted or misleading data, the agent may accept it without verification and incorporate it into its decisions. It is also becoming increasingly reliant on third-part tools and plugins adds to this exposure. If these components are compromised, they can affect behaviour without directly attacking the core system, creating a supply-side risk.
At this stage, the agent effectively operates as an insider. It holds legitimate credentials and interacts with systems in expected ways, making misuse harder to identify.
Application and Integration: System-Level Exposure
Agentic systems rarely operate in isolation. They are embedded in larger environments, interacting with identity systems, business logic, and operational workflows.
Access control becomes a major vulnerability. Agents tend to operate across multiple systems with various permission models, creating irregularities that can be exploited. Risks also arise from identity and delegation. In case an agent is operating on behalf of a user, then any vulnerabilities in authentication or session management can allow attackers to assume that authority.
Workflow execution amplifies these risks. Agents can initiate multi-step processes such as transactions, updates, or approvals. Manipulating a single step can change the result of the entire workflow. As integrations increase, so do the number of interaction points, making cumulative risk harder to track.
At this layer, failures are not isolated. They propagate into business operations, making consequences harder to contain.
Output and Action: Where Failures Become Visible
The output layer is where failures become visible, though they rarely originate there.
Data leakage has been a key concern. Agents may disclose information they are allowed to access, especially when tasks boundaries are not clearly defined. Misinformation and unsafe outputs are also important, particularly when outputs directly influence actions or decisions.
Generated code and commands introduce execution risk. If outputs are used without validation, errors or manipulations can have system-level effects. The shift towards autonomous action increases this risk, as small upstream deviations can lead to significant consequences without human intervention. This layer reflects symptoms rather than root causes. Addressing it alone does not reduce the underlying risk.
Beyond Layers: The Missing Dimension
A layered view helps, but it does not capture the full picture. Agentic systems are defined by continuous interaction across layers.
The key missing dimension is the runtime loop. Inputs shape reasoning, reasoning drives action, and actions feed back into both reasoning and memory. These cycles create feedback loops, where small manipulations may escalate over time. This also reduces observability. With multiple interacting components, it becomes difficult to trace cause and effect or identify where failures originate.
Supply chain dependencies add another layer of risk. Models, datasets, APIs, and plugins each introduce their own points of failure. A compromise at any of these points can propagate across the system. The attack surface also includes governance. Weak supervision, unclear responsibility, or excessive autonomy increase overall risk. Human control is not external to the system; it is part of its security.
Conclusion: Structuring the Attack Surface
Agentic AI expands the attack surface beyond traditional systems. It is both recursive and stateful. Risk does not just accumulate across layers; it moves and changes as the system operates.
Any useful representation must go beyond a linear stack. It should capture feedback loops, persistent state, and cross-layer dependencies that characterise the way these systems actually behave. The system is not a pipeline but a cycle. That is where both its capability and its risk emerge.
.webp)
Introduction
A Pew Research Center survey conducted in September 2023, found that among 1,453 age group of 13-17 year olds projected that the majority of the age group uses TikTok (63%), Snapchat (60%) and Instagram (59%) in the U.S. Further, in India the 13-19 year-olds age group makes up 31% of social media users in India, according to a report by Statista from 2021. This has been the leading cause of young users inadvertently or deliberately accessing adult content on social media platforms.
Brief Analysis of Meta’s Proposed AI Age Classifier
It can be seen as a step towards safer and moderated content for teen users, by placing age restrictions on teen social media users as sometimes they do not have enough cognitive skills to understand what content can be shared and consumed on these platforms and what can not as per their age. Moreover, there needs to be an understanding of platform policies and they need to understand that nothing can be completely erased from the internet.
Unrestricted access to social media exposes teens to potentially harmful or inappropriate online content, raising concerns about their safety and mental well-being. Meta's recent measures aim to address this, however striking a balance between engagement, protection, and privacy is also an essential part.
The AI-based Age Classifier proposed by Meta classifies users based on their age and places them in the ‘Teen Account’ category which has built-in limits on who can contact them, the content they see and more ways to connect and explore their interests. According to Meta, teens under 16 years of age will need parental permission to change these settings.
Meta's Proposed Solution: AI-Powered Age Classifier
This tool uses Artificial Intelligence (AI) to analyze users’ online behaviours and other profile information to estimate their age. It analyses different factors such as who follows the user, what kind of content they interact with, and even comments like birthday posts from friends. If the classifier detects that a user is likely under 18 years old, it will automatically switch them to a “Teen Account.” These accounts have more restricted privacy settings, such as limiting who can message the user and filtering the type of content they can see.
The adult classifier is anticipated to be deployed by next year and will start scanning for such users who may have lied about their age. All users found to be under 18 years old will be placed in the category of teen accounts, but 16-17 year olds will be able to adjust these settings if they want more flexibility, while younger teens will need parental permission. The effort is part of a broader strategy to protect teens from potentially harmful content on social media. This is especially important in today’s time as the invasion of privacy for anyone, particularly, can be penalised due to legal instruments like GDPR, DPDP Act, COPPA and many more.
Policy Implications and Compliances
Meta's AI Age Classifier addresses the growing concerns over teen safety on social media by categorizing users based on age, restricting minors' access to adult content, and enforcing parental controls. However, reliance on behavioural tracking might potentially impact the online privacy of teen users. Hence the approach of Meta needs to be aligned with applicable jurisdictional laws. In India, the recently enacted DPDP Act, of 2023 prohibits behavioural tracking and targeted advertising to children. Accuracy and privacy are the two main concerns that Meta should anticipate when they roll out the classifier.
Meta emphasises transparency to build user trust, and customizable parental controls empower families to manage teens' online experiences. This initiative reflects Meta's commitment to creating a safer, regulated digital space for young users worldwide, it must also align its policies properly with the regional policy and law standards. Meta’s proposed AI Age Classifier aims to protect teens from adult content, reassure parents by allowing them to curate acceptable content, and enhance platform integrity by ensuring a safer environment for teen users on Instagram.
Conclusion
Meta’s AI Age Classifier while promising to enhance teen safety and putting certain restrictions and parental controls on accounts categorised as ‘teen accounts’, must also properly align with global regulations like GDPR, and the DPDP Act with reference to India. This tool offers reassurance to parents and aims to foster a safer social media environment for teens. To support accurate age estimation and transparency, policy should focus on refining AI methods to minimise errors and ensure clear disclosures about data handling. Collaborative international standards are essential as privacy laws evolve. Meta’s initiative is intended to prioritise youth protection and build public trust in AI-driven moderation across social platforms, while it must also balance the online privacy of users while utilising these advanced tech measures on the platforms.
References
- https://familycenter.meta.com/in/our-products/instagram/
- https://www.indiatoday.in/technology/news/story/instagram-will-now-take-help-of-ai-to-check-if-kids-are-lying-about-their-age-on-app-2628464-2024-11-05
- https://www.bloomberg.com/news/articles/2024-11-04/instagram-plans-to-use-ai-to-catch-teens-lying-about-age
- https://tech.facebook.com/artificial-intelligence/2022/6/adult-classifier/
- https://indianexpress.com/article/technology/artificial-intelligence/too-young-to-use-instagram-metas-ai-classifier-could-help-catch-teens-lying-about-their-age-9658555/
