#FactCheck: Misleading Claim Amid West Asia Conflict: Old Yemen Video Shared as Iran’s Attack on Tel Aviv

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.

‍

Introduction

‍

The way we interact, go about our daily lives and manage our financial resources has completely changed in this digital age. Tasks that were earlier done manually (and were extremely time-consuming) now happen in just minutes and seconds. As this convenience expands, so do the risks that come with it. The proliferation of digital technology has opened up a wide arena where we, as users, are getting exposed to a complex environment of emerging cyber threats at every step. This has become especially true for our senior citizens, who face heightened vulnerabilities owing to their age and situational factors. Therefore, getting an understanding of these risks and knowing how to respond to them is no longer optional.

‍

Cybercrime against senior citizens is rising every year. The elderly today are using and depending on smartphones, net banking, UPI (Unified Payments Interface) and social media more than ever. Scammers are preying on this circumstance and perpetrating well-planned cyber attacks on the senior population. A 2024 report from the United States Federal Trade Commission (FTC) that analysed cybercrime data for the year 2023 found that older adults lost more money per incident as opposed to the younger population. 

‍

This blog aims at providing practical and easy-to-follow cyber safety tips for senior citizens and a clear action plan on how to respond if they ever fall victim to a scam.

‍

Safe and Smart Browsing for the Elderly: Essential Cybersecurity Tips

‍

1. Device safety: Cybersecurity can be ensured with one simple rule, which is securing your device first. This is the first line of defence against cybercrime and can go a long way in preventing online attacks and scams. This device security can be established by using screen locks such as setting up a PIN, lock patterns, fingerprints, or even applying the face lock feature on smartphones and tablets. 

‍

The next step is to install a reputable antivirus on your device (such as laptops and desktops) and keep updating it from time to time. Enabling additional security features in your devices, such as ‘automatic updates’, further ensures that the phone’s Operating System (OS) and the Applications (apps) remain up to date. This guarantees that the device’s OS and the apps are fully upgraded to catch and fix any new security threats. Users should also beware of any unauthorised software. Suspicious links or messages that offer free software or updates should be avoided at all costs. 

‍

The Government of India’s 2024 cyber safety booklet also stresses the importance of keeping devices and apps equipped with the latest security patches to prevent exploitation of any known vulnerabilities. 

‍

2. Robust Passwords and 2FA: Weak passwords are one of the easiest entry points for cybercriminals. This is why it is crucial to have strong and robust passwords to keep accounts and devices absolutely secure. It is also important to make sure to avoid using very basic PINs, such as date of birth or predictable ascending digits such as ‘12345’. Also, similar passwords should not be used across various bank, email or social media accounts. While creating a strong password, ideally, one should use a combination of upper case and lower case letters along with mixed digits and special characters. If need be, one can jot down these complex passwords in a separate physical diary to keep a record and refer to it in case of forgetfulness. 

‍

Enabling two-factor authentication adds an extra layer of protection. Whenever and wherever possible, users should enable OTP (one-time password) based verification or app-based authentication codes. In case someone does end up stealing the user’s password, they would still need the second factor to log in.

‍

The Government of India supported the Information Security Education and Awareness (ISEA) initiative, which also provides guidelines that recommend the use of passwords and secure authentication as key cyber safety practices for senior citizens.

‍

3. Handle messages with caution: Most cybercrime instances against senior citizens are initiated with a phone call, an SMS or a WhatsApp message. Personal details such as OTPs, CVV (3-digit number on the back of a debit/credit card), ATM or UPI PIN, full debit/credit card number, net banking ID and password, etc., should never be shared over a call or a chat. Also, people need to be aware of the fact that genuine officials will never make any demands for such details over calls or messages. ISEA has specifically warned about frauds like phishing, where fraudsters call or message their victims pretending to be from banks, courier companies, telecom operators or from the government to trick seniors into revealing their personal details.

‍

A caller might say ‘Your KYC can get blocked, give OTP now’ or ‘Your card will get deactivated, click this link’. Hang up immediately! Do not click any links, do not share any number and do not install any app that they suggest (eg. remote access apps).

‍

4. Banking without worries: Online banking is truly a boon, especially for senior citizens. It eliminates the need to visit the bank in person for every small requirement. But this comfort and ease can serve us well only if we use it with a certain sense of vigilance and responsibility. 

‍

Users are advised to use only official banking apps that are downloaded from Google Play Store or the Apple App Store. It is also prudent to activate SMS or email alerts for all online transactions so that the user receives timely notifications of any withdrawal or transfer. Users should also avoid using public Wi-Fi connections for banking or UPI payments. It is better  to use your own mobile data or a home Wi-Fi with a strong password. This safeguards your financial transactions. India has a system in place called the Citizen Financial Cyber Fraud Reporting and Management System (CFCFRMS) that quickly responds to online financial fraud, especially with regards to UPI and net banking frauds.

‍

5. Think before you share: Beyond online banking and other essential activities on the web, lies the world of social media which has seamlessly woven itself into our daily routines. Therefore, gaining awareness about the safe usage of social media is extremely important. For starters, users should keep the privacy settings of their social media accounts as ‘Private’. They should also be mindful of accepting friend requests from strangers or unknown profiles, even if they claim that they know the user. Users must also avoid ‘oversharing’ on public platforms and social media accounts. It is always best to refrain from posting about personal details on social media such as finances, travel schedules or addresses. If a user receives urgent messages from a close friend or a relative asking for money, then it is better to first call them up directly to verify that the request is genuine. 

‍

The older generation should also be wary of fraudsters who, under false identities, build emotional relationships online and then start demanding money, gifts or even private photos from their victims. 

‍

6. Stay a step ahead: Lastly, senior users can stay a step ahead in this game by familiarising themselves with the typical scams that are targeted towards their age group. This know-how will prepare them to identify scam patterns and early warning signs. Discussing experiences or sharing knowledge with family members or peers can also help seniors bridge the gap between awareness and action.

Fallen for a scam? Don’t panic! Here’s what seniors should do next: An action plan for recovery and protection

‍

We are all human at the end of the day. Anyone can fall for a scam, even the most careful people. The important thing is how to ‘respond’ after getting scammed. As per the 2023 Internet Crime Complaint Center (IC3) Elder Fraud Report, prompt reporting after a scam can improve the chances of either freezing or retrieving the stolen money. If you have fallen for a scam, here is what you should do:

‍

1. Keep calm and disconnect: Though it may feel challenging, but the first instinct that victims should display is to keep their calm and hang up the phone. The victim should then block the number and refrain from responding to any further emails or messages. Any remote access app, screen sharing app or a support tool that were installed at the request of the caller, should be instantly uninstalled.

‍

2. Alert your bank right away: Time is of the essence when it comes to online financial frauds. If a user has been scammed and he is aware of it, then he should immediately call the official customer care number of his bank. It is to be noted here that this number should be verified from the bank’s website or from the user’s credit/debit card. Once the user is connected to the bank’s support staff, he should inform them clearly that he has been a victim of an online fraud. The support staff may guide the user regarding the options that he can undertake. These can be: blocking the cards, freezing the account or an attempt from the bank’s interface to try and stop or recall the recent transaction. 

‍

The Reserve Bank of India (RBI) has laid down certain guidelines on unauthorised electronic banking transactions that encourage quick reporting and provide a steady framework for customer liability. Also, if in case a victim feels that the bank’s response has been unsatisfactory, the complaints can be further escalated to the RBI Ombudsman.

‍

3. Reset online credentials: Once the victim has spoken to the bank and taken all necessary steps in that regard, the next step is to reset the passwords. Passwords for netbanking, UPI apps, email accounts, social media, shopping platforms, etc., should all be changed. The victim can also turn on two factor authentication for all important accounts. Incase some critical information like PAN (Permanent Account Number) or Aadhar or SSN (Social Security Number ) has been shared, the victim should be on the lookout for any identity theft in the form of unknown loans, new cards or credit enquiries. 

‍

4. Notify official cybercrime portals or authorities: The next step is to inform the relevant cybercrime authorities in your region. In India there is a government backed National Cyber Crime Helpline which is a part of the Financial Cyber Fraud Reporting and Management System. The aim of the helpline is to prevent further losses in digital payment frauds. Victims can also file a complaint on the National Cyber Crime Reporting portal. Other than these options, the victim can visit the nearest police station along with copies of their online complaint, bank statements, transaction receipts and ID proofs. 

‍

In the US, such complaints can be lodged with the IC3 which runs a specialised elder fraud program. The IC3 also publishes the annual Elder Fraud Report and lays down latest guidelines and trends regarding cybercrime.

‍

5. Preserve proofs of the scam: The IC3’s 2023 report notes that detailed and prompt reporting by the victim enables law enforcement and financial institutions to deploy ‘Financial Fraud Kill Chain’ procedures that help in tracing or freezing stolen funds. This is why it is very important to preserve the evidence of a cyber scam. Proper documentation can help banks and law enforcement agencies to properly investigate a case. If possible, the victim should save all messages, WhatsApp chats, emails, screenshots and voice recordings. Establishing a clear timeline of the events also helps. Minute details such as: when the first call was received, what was said and when the money exited the victim’s account, significantly help in piecing together the full picture. Victims should also keep a record of all bank statements and transaction alert messages related to the fraudulent activity. 

‍

6. Helping seniors heal: Apart from monetary damage, the emotional impact of a scam should also be acknowledged. Getting scammed can be deeply humiliating for senior citizens, especially at their age where everyday life may already feel quite overwhelming. Also, when it comes to their lifelong savings, the ramifications of becoming a victim of an online financial scam can be quite distressing for seniors. The US Department of Justice’s elder fraud content emphasises on the fact that senior victims need ample emotional support and should not be blamed or shamed for their predicament.

‍

Families and caregivers can guide the elderly victims step by step through the recovery process which includes taking the necessary technical steps, filling out forms and following up with banks or authorities. For large value frauds, repeated targeting or cases that involve identity theft, getting a legal consult for the victim is highly advisable. This kind of support and reassurance can help seniors reduce their anxiety and regain their inner strength.

‍

Secure today, safe tomorrow

‍

The rate of cybercrime against senior citizens isn’t showing any signs of slowing down. Scammers are refining their techniques every day. Therefore, the best long standing defence against cybercrime is to make cyber safety a regular habit. This can be achieved with the support of families, caregivers and communities. Beyond the practical steps, spreading awareness and maintaining an open dialogue is equally important. Senior citizens should feel comfortable asking questions, sharing concerns and continue learning from experiences, be it theirs or of others. Reinforcement of safe online practices should become ingrained at the core of every society. When cybersecurity practices merge with everyday life, seniors gain both protection and confidence. Simple and steady digital safety practices that come with guidance and reassurance empower the seniors to enjoy the benefits of technology without fear.

Simply put, the habits that are adopted today will lay the foundation for a secure and safer digital tomorrow.

‍

References  

• https://bankingjournal.aba.com/2024/11/ftc-older-adults-lost-up-to-61-5b-to-fraud-in-2023/
• https://www.cert-in.org.in/PDF/CSH_Booklet.pdf
• https://infosecawareness.in/concept/cyber-security-tips-for-senior-citizens
• https://cybercrime.gov.in/UploadMedia/instructions_citizenreportingcyberfrauds.pdf
• https://www.ic3.gov/annualreport/reports/2023_ic3elderfraudreport.pdf
• https://www.arklegal.in/post/how-to-report-cyber-crime-in-india
• https://www.pib.gov.in/Pressreleaseshare.aspx?PRID=1814120
• https://cybercrime.gov.in/Webform/crmcondi.aspx
• https://www.justice.gov/archives/stopfraud-archive/elder-fraud-and-financial-exploitation
• https://www.ftc.gov/system/files/ftc_gov/pdf/federal-trade-commission-protecting-older-adults-report_102024.pdf
• https://www.unionbankofindia.bank.in/pdf/cyber-security-customer-awareness-guide-vol-v.pdf
• https://cdnbbsr.s3waas.gov.in/s371e09b16e21f7b6919bbfc43f6a5b2f0/uploads/2024/11/20241111752149945.pdf
• https://www.dsci.in/files/content/documents/2024/CSAM24-Infographic_Cyber-Safety-Essentials-for-Senior-Citizens-v1.pdf

‍

Executive Summary

Claims are circulating that Iran’s Supreme Leader Ayatollah Ali Khamenei was killed in a major attack allegedly carried out by Israel and the United States. Amid these claims, a video is being widely shared on social media in which Khamenei can be heard saying, “Beware of fake news, I am alive.” Research conducted by CyberPeace has found the viral claim to be false. Our research revealed that the video being shared is old and that Khamenei’s voice has been altered using artificial intelligence to support a misleading narrative.

‍

Claim

‍

On March 1, 2026, an Instagram user shared the viral video in which Ayatollah Ali Khamenei is heard saying, “Beware of fake news, I am alive.” The link to the post and its archived version are provided above along with a screenshot.

• https://www.instagram.com/reel/DVWWA9vkU-6/ 
• https://archive.ph/g30Ce 

‍

‍

Fact Check:

‍

To verify the authenticity of the claim, we extracted key frames from the viral video and conducted a reverse image search using Google Lens. During the research, we found the same video on the YouTube channel of Sky News Australia, published on June 19, 2025. In the approximately 43-minute-long video, the portion used in the viral clip appears around the 10-minute mark.

• https://www.youtube.com/watch?v=9mmZblh3blo 

‍

‍

According to Sky News Australia’s report, Iran’s Supreme Leader Ayatollah Ali Khamenei had rejected US President Donald Trump’s demand for unconditional surrender. The Ayatollah regime also warned that any American military intervention would be accompanied by “irreparable damage.” Upon closely listening to the viral clip, we noticed that Khamenei’s voice sounded robotic, raising suspicion that it may have been AI-generated. We then analyzed the video using the AI detection tool AURGIN AI. The results indicated that the viral clip had been generated using artificial intelligence.

‍

‍

Conclusion

‍

Our research establishes that the viral video is old and has been digitally manipulated. Ayatollah Ali Khamenei’s voice has been altered using artificial intelligence and the clip is being shared with a misleading claim.

‍