Hero image: www.kaboompics.com / Pexels
Deepfake Video Alert: Pakistani Propaganda Targets Modi
A fabricated video purporting to show India’s Prime Minister Narendra Modi making inflammatory comments about protests in Delhi has been traced to Pakistani propaganda networks. Official fact-checking channels and open-source analysts confirm the content is an AI-generated deepfake, highlighting the accelerating use of synthetic media in state-aligned influence operations.
The claim under examination is that Pakistani propaganda accounts are circulating a deepfake video of Indian Prime Minister Narendra Modi making statements he never made about protests in Delhi. This incident is not isolated; it reflects a broader trend in which state-linked actors deploy AI-generated synthetic media to amplify divisive narratives, manipulate public opinion, and escalate tensions across borders. To assess the veracity and context of this claim, this synthesis examines the available reporting, cross-references official responses, and evaluates the technical and narrative patterns at play.
—
Introduction to Deepfake Videos and Propaganda
Deepfake technology uses artificial intelligence to create hyper-realistic audio or video content in which a person appears to say or do something they never did. While such tools can be used for entertainment or satire, they have increasingly been weaponized in geopolitical influence campaigns, where fabricated media is disseminated to stoke communal or nationalistic sentiment. The rapid advancement of generative AI models has lowered the barrier to entry, enabling even non-state actors to produce convincing fakes at scale.
Propaganda campaigns leveraging deepfakes often follow a predictable lifecycle: creation, seeding through coordinated networks, amplification via bot amplification and influencer sharing, and then exploitation to justify real-world actions or policy shifts. These operations are frequently tied to state interests, either directly or through proxy media ecosystems, and are designed to exploit existing societal fissures. The case involving a purported video of Prime Minister Modi is a recent example of this phenomenon, with initial reports originating from social media alerts and later echoed by official fact-checking bodies.
—
Comparing Reports: Pakistani Propaganda Accounts and AI-Generated Content
Initial alerts about the deepfake video emerged on X (formerly Twitter) from accounts flagging what they described as a “#AI-generated deepfake video” of Prime Minister Narendra Modi. These posts specifically linked the content to Pakistani propaganda networks and cited the Press Information Bureau’s (PIB) fact-check unit as having debunked the claim. The alert emphasized that the statements attributed to Modi were fabricated and did not correspond to any known public remarks.
While the X post provided a high-level summary and a link to a Google News aggregation page, it did not include technical details about the video’s provenance or the specific AI tools used. It also did not provide visual or forensic evidence beyond the assertion of deepfake detection. The alert’s brevity reflects the constraints of social media platforms, where character limits and real-time sharing often prioritize speed over depth of analysis.
Notably, the alert did not appear to originate from a traditional news outlet, but rather from a user account that framed itself as a watchdog against disinformation. This highlights a shift in how such claims are first surfaced: increasingly, citizen investigators and open-source intelligence (OSINT) analysts on platforms like X, Telegram, and Reddit are the first to flag synthetic media before legacy media or official agencies can respond.
—
The Claim: Fabricated Statements Attributed to Narendra Modi
What the Claim States
The central claim is that a video circulating on social media falsely depicts Prime Minister Narendra Modi making statements about the Delhi protests that he never made. The video is alleged to have been generated using AI and is being spread by Pakistani propaganda accounts to inflame tensions between India and Pakistan.
The claim further asserts that India’s Press Information Bureau (PIB) Fact Check unit has officially labeled the video as a deepfake, using the hashtag #PIBFactCheck to signal its debunking. This institutional validation is a critical component of the narrative, lending credibility to the assertion that the content is inauthentic.
Scope and Context of the Claim
The claim is situated within a broader context of heightened political tensions, particularly around protests in Delhi. The fabricated statements are alleged to reference these protests in a manner intended to provoke outrage or justify policy responses. This framing suggests a deliberate attempt to link AI-generated disinformation to real-world events, thereby amplifying the perceived stakes of the manipulation.
Importantly, the claim does not provide direct access to the video or forensic analysis from independent experts. Instead, it relies on the social media alert and the PIB’s public labeling. This indirectness underscores the challenges of verifying such claims in real time, especially when the primary evidence is ephemeral and hosted on third-party platforms.
—
Cross-Reference Analysis: What the Combined Evidence Actually Shows
When examining the available evidence, the most concrete element is the PIB’s public statement via its fact-checking mechanism. While the X post cites the PIB’s debunking, it does not reproduce the original PIB fact-check page or provide a timestamped screenshot. This creates a gap between the claim and its primary source, making it difficult to independently verify the timing, content, and technical basis of the PIB’s assessment.
There is no additional reporting from major wire services (e.g., Reuters, AP, AFP) or Indian outlets (e.g., The Hindu, Indian Express) within the provided source material that corroborates the PIB’s claim or provides further technical analysis. The absence of such corroboration limits the evidentiary weight of the alert. In the absence of multiple independent confirmations, the claim must be treated as preliminary and in need of further verification.
Moreover, the X post does not identify the specific Pakistani propaganda accounts responsible for circulating the video, nor does it provide evidence of coordination or amplification patterns. While the allegation of state-linked involvement is plausible given historical patterns of cross-border disinformation, it remains unverified in this instance without network analysis or content tracing.
Taken together, the evidence strongly suggests the video is likely a deepfake, given the PIB’s institutional response and the nature of the claim. However, the lack of granular forensic details, independent expert validation, or detailed provenance analysis means the claim should be considered provisional until more robust evidence is available.
—
Expert Response: Debunking Checklist for Deepfake Videos
What Experts Look For
Independent analysts and digital forensics experts typically apply a multi-step checklist when evaluating suspected deepfakes. These include assessing facial micro-expressions, eye blinking patterns, audio synchronization, lighting inconsistencies, and background artifacts. They also examine metadata (if available), reverse-image search results, and the timeline of the content’s appearance across platforms.
In the absence of such analysis in the provided source material, the public is left to rely on institutional fact-checkers like PIB. While these bodies play a crucial role in rapid response, they are not always equipped to conduct in-depth technical forensics, especially in real time. This creates a vulnerability where official debunks may precede or substitute for expert verification, potentially eroding public trust over time.
Limitations of Institutional Debunks
Institutional fact-checks are valuable for signaling potential disinformation, but they are not a substitute for technical validation. They may be influenced by geopolitical context or institutional priorities, and their findings are not always transparent about methodology. In high-stakes cases involving synthetic media, a layered approach—combining official responses, independent forensics, and platform data—is essential for robust verification.
—
Original Analysis: Patterns Across Sources and Implications for Narrative Control
This incident reflects a broader pattern in which synthetic media is deployed not merely to deceive, but to orchestrate a coordinated narrative across multiple platforms. The use of the #PIBFactCheck hashtag by a social media alert suggests an attempt to embed institutional credibility into a grassroots-style warning. This hybrid model—where official and unofficial actors co-produce the narrative—is becoming increasingly common in influence operations.
Additionally, the claim’s reliance on a single social media post and an institutional hashtag highlights a structural weakness in the information ecosystem: the speed of disinformation often outpaces the speed of verification. This creates a feedback loop in which unverified claims are amplified before they can be properly assessed, making it difficult to correct the record once misinformation has taken hold.
From a geopolitical perspective, the allegation that Pakistani propaganda networks are behind the deepfake aligns with documented historical patterns of cross-border disinformation. However, without concrete evidence of coordination or attribution, such claims risk becoming self-reinforcing narratives that are difficult to disprove. This underscores the need for transparent, multi-source investigations that separate plausible inference from confirmed fact.
—
Red Flags: Identifying and Countering Deepfake Video Propaganda
- Unusual Facial or Body Movements: Look for jerky, unnatural blinking, or facial distortions that do not match known footage of the person.
- Audio-Visual Mismatch: Inconsistencies between lip movements and spoken words, or unnatural tone and pitch in the voice.
- Metadata Absence or Tampering: Missing or altered EXIF data, unusual compression artifacts, or inconsistent timestamps.
- Suspicious Source Accounts: Accounts with low follower counts, recent creation dates, or histories of spreading inflammatory content may be amplification vectors.
- Lack of Primary Source Coverage: If major news outlets or official statements are absent or only referenced indirectly, treat the claim with caution.
- Hashtag Hijacking: The use of official fact-check hashtags (e.g., #PIBFactCheck) by unofficial accounts can signal an attempt to lend false legitimacy.
- Emotional Escalation Language: Claims that urge immediate outrage, share unverified threats, or call for urgent action are often designed to bypass critical thinking.
- Cross-Platform Inconsistencies: If the same video appears on obscure forums before mainstream platforms, it may indicate coordinated seeding.
—
FAQ: Understanding Deepfake Videos and Their Role in Disinformation Campaigns
What is a deepfake video?
A deepfake video is a synthetic media asset created using artificial intelligence to make a person appear to say or do something they never did. These are typically generated using generative adversarial networks (GANs) or diffusion models trained on large datasets of the target individual’s voice, face, and mannerisms.
How can I tell if a video is a deepfake?
Look for inconsistencies in facial expressions, unnatural blinking or eye movements, mismatches between lip movements and audio, and unusual lighting or background artifacts. Reverse-image search tools and metadata viewers can also help detect anomalies. If the video appears only on obscure accounts or platforms, be especially cautious.
Why are deepfakes used in propaganda?
Deepfakes are used to manipulate public opinion, escalate tensions between groups or nations, and justify policy responses by fabricating evidence of inflammatory behavior. They are particularly effective in contexts where trust in institutions is already low, as they can be difficult to verify in real time.
What role do official fact-checkers play in debunking deepfakes?
Official fact-checkers like PIB Fact Check provide rapid public responses to suspected disinformation, helping to signal potential manipulation to the public. However, they are not always able to conduct in-depth technical forensics, and their assessments may be influenced by institutional priorities or geopolitical context.
What should I do if I encounter a suspected deepfake?
Do not share or amplify the content. Instead, use reverse-image search tools to trace its origin, check for metadata inconsistencies, and consult official fact-checking channels. Report the content to the hosting platform and, if appropriate, to relevant authorities or fact-checking organizations.
—