How pro-BJP accounts are using AI videos of Dhruv Rathee to push political messaging

Hyderabad: YouTuber and political analyst Dhruv Rathee is widely known for his criticism of the government. For many who follow his videos, his explainer videos often ‘question authority’ even as BJP supporters label him ‘anti-Modi’ or ‘anti-BJP’.

Viral videos show Rathee praising PM Modi

However, a series of videos circulating on social media now appear to show Rathee praising Prime Minister Narendra Modi and the BJP government. These clips are being widely shared by pro-BJP accounts without any indication or warning that they may be manipulated or artificially generated.

What is the content of the videos?

A Twitter account named ‘Hindutva Knight’ shared a video of Rathee in which he appears to say that the PM Modi-led government had reduced the LPG crisis even before the shortage began. (Archive)

We reviewed the account and found that its posts are largely pro-BJP and critical of opposition parties.

Another Instagram account named ‘politics humour’, which posts pro-BJP content and criticises other political parties, posted a similar video of Rathee. In this video, he praises the current Central government for transforming Indian Railways and criticises the previous UPA government. (Archive)

An Instagram handle named ‘n4mobharat.in Narendra Modi | Prime Minister of Bharat | NamoBharat’—who identifies themselves as a fan of Modi in the bio—posted seven such videos of Dhurv Rathee, where he appears to praise PM Modi or defend the BJP government.

But are these videos featuring Dhurv Rathee real? No, they are created using Artificial Intelligence (AI).

We further found that these AI-generated videos have better engagement than videos with real content shared from the handles.

In one of the videos, Rathee is seen advocating for a double-engine government and giving West Bengal as an example of why two different governments do not work.

One more Instagram account named ‘pmmodi.again | Himanshu Kumawat’ published the same videos. The handle has an image of PM Modi and Uttar Pradesh Chief Minister Yogi Adityanath as its display picture. The bio reads “Fan Page of @narendramodi Ji.”

The account’s feed looks identical to the Instagram handle mentioned earlier, as the content posted is the same.

How did we find out that they are AI videos?

The videos follow a common pattern when it comes to the presentation of content.

They start with Rathee explaining a topic while screenshots of newspaper reports and tweets posted by media outlets are shown as a way to show that the explanation is based on ‘facts’.

The content in these videos is mainly targeting the Opposition and praising the PM Modi-led government while claiming to show ‘evidence’. Notwithstanding whether the reports shown in the video are real or fake, the claim that the video really features Dhruv Rathee speaking is false.

What does the analysis reveal?

Upon comparing the video using existing footage, we could determine that the videos were not created by manipulating actual footage, where the video remains the same, but the audio is different.

Unlike typical deepfake videos, where footage of prominent personalities is taken from real speeches and then manipulated using AI, these clips have no identifiable source.

In most deepfake cases, the original video can be traced and compared to show clear manipulation. However, no such reference exists for these clips. They look a bit similar to videos of Dhruv Rathee but they are not exactly the same.

Instead, these videos appear to have been generated entirely from scratch using prompts on generative AI platforms. This likely explains why the visuals are far more inconsistent than standard deepfakes, as there is no original reference footage guiding the output.

Visual discrepancies

The physical movement of the person in the video does not match that of Dhruv Rathee. The movements look awkward, and his teeth appear brighter and wider compared to the person.

AI-generation tools often struggle with creating believable teeth and lip sync and the videos are a good example of that. The hand movements are robotic and the fingers look inconsistent. The voice resembles Rathee’s voice but does not sound human. It takes pauses and mispronounces words as AI models do when generating fake audio.

The videos are poorly made AI-generated content. They come under the category of easily identifiable AI footage. However, the line remains blurry for those who are yet to understand the difference between AI and real content.

AI detection tools flag manipulation

We uploaded the videos to the Hive Moderation AI detector. The tool found that both the video and audio are likely AI-generated or deepfake content with probability scores above 90 per cent in all cases.

We also analysed the clips using Google Gemini. The tool noted that the videos may not have been created using its own AI systems but indicated that they are AI-generated.

Gemini further added that videos featuring these types of digital avatars are typically created using dedicated video generation platforms such as Synthesia, HeyGen or D-ID, which specialise in ‘text-to-video’ digital twins.

Pattern suggests a coordinated campaign

These videos suggest a pattern where self-proclaimed pro-BJP accounts are using AI-generated clips of Dhruv Rathee to push a specific political message. Dhruv Rathee is a popular online personality for his contentious political content. He has more than 31 million subscribers.

By making him appear to praise the PM Modi-led government, the videos can mislead his viewers who may not realise that the content is not real, or at least, manipulated.

Even though different accounts have shared these videos on various platforms, they appear to be made using the same AI model, as they look identical. This suggests a coordinated effort as the content gains traction and spreads the message.

How are these videos different?

Previously, we have seen deepfakes of PM Modi, Finance Minister Nirmala Sitharaman and actor Rashmika Mandanna, and they were used to spread false claims and run scams. NewsMeter has previously debunked several such videos.

These videos are different because the speeches appear more opinionated. For example, in one clip, Rathee seems to reiterate the government’s statements on fuel shortages and slam the Opposition for spreading false information. In another clip, he appears to target the Congress over an incident in which a group of Muslim youths were allegedly arrested for organising a fast-breaking gathering during the month of Ramadan in Varanasi.

The manipulation lies in how the message is framed and who is shown delivering it. Dhruv Rathee’s likeness is used to present information in a way that portrays the BJP government’s actions positively while targeting Opposition parties. This makes the videos harder to detect and more effective at influencing viewers.

How do Indians fare in distinguishing between real and fake?

In India, distinguishing between AI-generated and real content is still a challenge for many users.

A 2024 survey by the cybersecurity and online protection company McAfee found that nearly 75 per cent of Indians have encountered deepfake content online; 1 in 4 said they had come across political deepfakes that they later realised were fake.