#FactCheck -Fake Viral Poster Falsely Attributes Political Endorsement of Rahul Gandhi to Sachin Tendulkar

As AI language models become more powerful, they are also becoming more prone to errors. One increasingly prominent issue is AI hallucinations, instances where models generate outputs that are factually incorrect, nonsensical, or entirely fabricated, yet present them with complete confidence. Recently, ChatGPT released two new models—o3 and o4-mini, which differ from earlier versions as they focus more on step-by-step reasoning rather than simple text prediction. With the growing reliance on chatbots and generative models for everything from news summaries to legal advice, this phenomenon poses a serious threat to public trust, information accuracy, and decision-making.

What Are AI Hallucinations?

AI hallucinations occur when a model invents facts, misattributes quotes, or cites nonexistent sources. This is not a bug but a side effect of how Large Language Models (LLMs) work, and it is only the probability that can be reduced, not their occurrence altogether. Trained on vast internet data, these models predict what word is likely to come next in a sequence. They have no true understanding of the world or facts, they simulate reasoning based on statistical patterns in text. What is alarming is that the newer and more advanced models are producing more hallucinations, not fewer. seemingly counterintuitive. This has been prevalent reasoning-based models, which generate answers step-by-step in a chain-of-thought style. While this can improve performance on complex tasks, it also opens more room for errors at each step, especially when no factual retrieval or grounding is involved.

As per reports shared on TechCrunch, it mentioned that when users asked AI models for short answers, hallucinations increased by up to 30%. And a study published in eWeek found that ChatGPT hallucinated in 40% of tests involving domain-specific queries, such as medical and legal questions. This was not, however, limited to this particular Large Language Model, but also similar ones like DeepSeek. Even more concerning are hallucinations in multimodal models like those used for deepfakes. Forbes reports that some of these models produce synthetic media that not only look real but are also capable of contributing to fabricated narratives, raising the stakes for the spread of misinformation during elections, crises, and other instances.

It is also notable that AI models are continually improving with each version, focusing on reducing hallucinations and enhancing accuracy. New features, such as providing source links and citations, are being implemented to increase transparency and reliability in responses.

The Misinformation Dilemma

The rise of AI-generated hallucinations exacerbates the already severe problem of online misinformation. Hallucinated content can quickly spread across social platforms, get scraped into training datasets, and re-emerge in new generations of models, creating a dangerous feedback loop. However, it helps that the developers are already aware of such instances and are actively charting out ways in which we can reduce the probability of this error. Some of them are:

1. Retrieval-Augmented Generation (RAG): Instead of relying purely on a model’s internal knowledge, RAG allows the model to “look up” information from external databases or trusted sources during the generation process. This can significantly reduce hallucination rates by anchoring responses in verifiable data.
2. Use of smaller, more specialised language models: Lightweight models fine-tuned on specific domains, such as medical records or legal texts. They tend to hallucinate less because their scope is limited and better curated.

Furthermore, transparency mechanisms such as source citation, model disclaimers, and user feedback loops can help mitigate the impact of hallucinations. For instance, when a model generates a response, linking back to its source allows users to verify the claims made.

Conclusion

AI hallucinations are an intrinsic part of how generative models function today, and such a side-effect would continue to occur until foundational changes are made in how models are trained and deployed. For the time being, developers, companies, and users must approach AI-generated content with caution. LLMs are, fundamentally, word predictors, brilliant but fallible. Recognising their limitations is the first step in navigating the misinformation dilemma they pose.

References

• https://www.eweek.com/news/ai-hallucinations-increase/
• https://www.resilience.org/stories/2025-05-11/better-ai-has-more-hallucinations/ 
• https://www.ekathimerini.com/nytimes/1269076/ai-is-getting-more-powerful-but-its-hallucinations-are-getting-worse/ 
• https://techcrunch.com/2025/05/08/asking-chatbots-for-short-answers-can-increase-hallucinations-study-finds/ 
• https://en.as.com/latest_news/is-chatgpt-having-robot-dreams-ai-is-hallucinating-and-producing-incorrect-information-and-experts-dont-know-why-n/
• https://www.newscientist.com/article/2479545-ai-hallucinations-are-getting-worse-and-theyre-here-to-stay/
• https://www.forbes.com/sites/conormurray/2025/05/06/why-ai-hallucinations-are-worse-than-ever/ 
• https://towardsdatascience.com/how-i-deal-with-hallucinations-at-an-ai-startup-9fc4121295cc/ 
• https://www.informationweek.com/machine-learning-ai/getting-a-handle-on-ai-hallucinations 

‍

Executive Summary:

A video showing a car catching fire is rapidly going viral on social media. In the clip, a family can be seen bursting firecrackers in front of a newly purchased car. Moments later, the vehicle also appears to catch fire. The video is being shared with the claim that the family was celebrating the purchase of a new car with fireworks, which accidentally led to the vehicle going up in flames. Many users are circulating the clip as footage of a real incident. However, an research by the CyberPeace found that the video is not from a real-life event but has been created using Artificial Intelligence (AI).

‍

Claim

‍

On February 25, 2026, an X user named “Mamta Rajgarh” shared the viral video with the caption:“This was supposed to be a grand celebration for buying a new car, but it turned into a ceremony of burning the car. What do you say? Comment below.”

‍

• Post link: https://x.com/rajgarh_mamta1/status/2026696175311786408?s=20
• Archived link: https://perma.cc/22AA-KBS4

‍

‍

‍

Fact Check:

‍

To verify the claim, we conducted a keyword search on Google but found no credible news reports supporting the alleged incident. Upon closely examining the video, we noticed several technical inconsistencies. The car’s number plate is unclear, a common flaw often seen in AI-generated content. Additionally, the sequence of events appears unnatural — the firecrackers seem to extinguish first, and only after a delay does the car suddenly catch fire. These irregularities raised suspicion that the video may have been artificially generated. To further verify, we analyzed the clip using AI detection tools. Hive Moderation indicated a 98.7 percent likelihood that the video was generated using Artificial Intelligence.

‍

Another AI detection tool, Undetectable.ai, suggested a 77 percent probability that the video was AI-generated.

‍

Conclusion

‍

Our research confirms that the viral video does not depict a real incident. It has been created using Artificial Intelligence and is being misleadingly shared as genuine footage.

‍

Introduction

Deepfake technology, which combines the words "deep learning" and "fake," uses highly developed artificial intelligence—specifically, generative adversarial networks (GANs)—to produce computer-generated content that is remarkably lifelike, including audio and video recordings. Because it can provide credible false information, there are concerns about its misuse, including identity theft and the transmission of fake information. Cybercriminals leverage AI tools and technologies for malicious activities or for committing various cyber frauds. By such misuse of advanced technologies such as AI, deepfake, and voice clones. Such new cyber threats have emerged. 

India Topmost destination for deepfake attacks

According to Sumsub’s identity fraud report 2023, a well-known digital identity verification company with headquarters in the UK. India, Bangladesh, and Pakistan have become an important participants in the Asia-Pacific identity fraud scene with India’s fraud rate growing exponentially by 2.99% from 2022 to 2023. They are among the top ten nations most impacted by the use of deepfake technology. Deepfake technology is being used in a significant number of cybercrimes, according to the newly released Sumsub Identity Fraud Report for 2023, and this trend is expected to continue in the upcoming year. This highlights the need for increased cybersecurity awareness and safeguards as identity fraud poses an increasing concern in the area. 

How Deeepfake Works 

Deepfakes are a fascinating and worrisome phenomenon that have emerged in the modern digital landscape. These realistic-looking but wholly artificial videos have become quite popular in the last few months. Such realistic-looking, but wholly artificial, movies have been ingrained in the very fabric of our digital civilisation as we navigate its vast landscape. The consequences are enormous and the attraction is irresistible.  

Deep Learning Algorithms

Deepfakes examine large datasets, frequently pictures or videos of a target person, using deep learning techniques, especially Generative Adversarial Networks. By mimicking and learning from gestures, speech patterns, and facial expressions, these algorithms can extract valuable information from the data. By using sophisticated approaches, generative models create material that mixes seamlessly with the target context. Misuse of this technology, including the dissemination of false information, is a worry. Sophisticated detection techniques are becoming more and more necessary to separate real content from modified content as deepfake capabilities improve.

Generative Adversarial Networks

Deepfake technology is based on GANs, which use a dual-network design. Made up of a discriminator and a generator, they participate in an ongoing cycle of competition. The discriminator assesses how authentic the generated information is, whereas the generator aims to create fake material, such as realistic voice patterns or facial expressions. The process of creating and evaluating continuously leads to a persistent improvement in Deepfake's effectiveness over time. The whole deepfake production process gets better over time as the discriminator adjusts to become more perceptive and the generator adapts to produce more and more convincing content. 

 Effect on Community

The extensive use of Deepfake technology has serious ramifications for several industries. As technology develops, immediate action is required to appropriately manage its effects. And promoting ethical use of technologies. This includes strict laws and technological safeguards. Deepfakes are computer trickery that mimics prominent politicians' statements or videos. Thus, it's a serious issue since it has the potential to spread instability and make it difficult for the public to understand the true nature of politics. Deepfake technology has the potential to generate totally new characters or bring stars back to life for posthumous roles in the entertainment industry. It gets harder and harder to tell fake content from authentic content, which makes it simpler for hackers to trick people and businesses.

Ongoing Deepfake Assaults In India

Deepfake videos continue to target popular celebrities, Priyanka Chopra is the most recent victim of this unsettling trend. Priyanka's deepfake adopts a different strategy than other examples including actresses like Rashmika Mandanna, Katrina Kaif, Kajol, and Alia Bhatt. Rather than editing her face in contentious situations, the misleading film keeps her look the same but modifies her voice and replaces real interview quotes with made-up commercial phrases. The deceptive video shows Priyanka promoting a product and talking about her yearly salary, highlighting the worrying development of deepfake technology and its possible effects on prominent personalities.

Actions Considered by Authorities

A PIL was filed requesting the Delhi High Court that access to websites that produce deepfakes be blocked. The petitioner's attorney argued in court that the government should at the very least establish some guidelines to hold individuals accountable for their misuse of deepfake and AI technology. He also proposed that websites should be asked to identify information produced through AI as such and that they should be prevented from producing illegally. A division bench highlighted how complicated the problem is and suggested the government (Centre) to arrive at a balanced solution without infringing the right to freedom of speech and expression (internet).  

Information Technology Minister Ashwini Vaishnaw stated that new laws and guidelines would be implemented by the government to curb the dissemination of deepfake content. He presided over a meeting involving social media companies to talk about the problem of deepfakes. "We will begin drafting regulation immediately, and soon, we are going to have a fresh set of regulations for deepfakes. this might come in the way of amending the current framework or ushering in new rules, or a new law," he stated.

Prevention and Detection Techniques

To effectively combat the growing threat posed by the misuse of deepfake technology, people and institutions should place a high priority on developing critical thinking abilities, carefully examining visual and auditory cues for discrepancies, making use of tools like reverse image searches, keeping up with the latest developments in deepfake trends, and rigorously fact-check reputable media sources. Important actions to improve resistance against deepfake threats include putting in place strong security policies, integrating cutting-edge deepfake detection technologies, supporting the development of ethical AI, and encouraging candid communication and cooperation. We can all work together to effectively and mindfully manage the problems presented by deepfake technology by combining these tactics and adjusting the constantly changing terrain.

Conclusion

Advanced artificial intelligence-powered deepfake technology produces extraordinarily lifelike computer-generated information, raising both creative and moral questions. Misuse of tech or deepfake presents major difficulties such as identity theft and the propagation of misleading information, as demonstrated by examples in India, such as the latest deepfake video involving Priyanka Chopra. It is important to develop critical thinking abilities, use detection strategies including analyzing audio quality and facial expressions, and keep up with current trends in order to counter this danger. A thorough strategy that incorporates fact-checking, preventative tactics, and awareness-raising is necessary to protect against the negative effects of deepfake technology. Important actions to improve resistance against deepfake threats include putting in place strong security policies, integrating cutting-edge deepfake detection technologies, supporting the development of ethical AI, and encouraging candid communication and cooperation. We can all work together to effectively and mindfully manage the problems presented by deepfake technology by combining these tactics and making adjustments to the constantly changing terrain. Creating a true cyber-safe environment for netizens. 

References:

• https://yourstory.com/2023/11/unveiling-deepfake-technology-impact
• https://www.indiatoday.in/movies/celebrities/story/deepfake-alert-priyanka-chopra-falls-prey-after-rashmika-mandanna-katrina-kaif-and-alia-bhatt-2472293-2023-12-05
• https://www.csoonline.com/article/1251094/deepfakes-emerge-as-a-top-security-threat-ahead-of-the-2024-us-election.html
• https://timesofindia.indiatimes.com/city/delhi/hc-unwilling-to-step-in-to-curb-deepfakes-delhi-high-court/articleshow/105739942.cms
• https://www.indiatoday.in/india/story/india-among-top-targets-of-deepfake-identity-fraud-2472241-2023-12-05
• https://sumsub.com/fraud-report-2023/