#FactCheck - AI-Generated Video Falsely Shared as ‘Multi-Hooded Snake’ Sighting in Vrindavan

Executive Summary:

Several videos claiming to show bizarre, mutated animals with features such as seal's body and cow's head have gone viral on social media. Upon thorough investigation, these claims were debunked and found to be false. No credible source of such creatures was found and closer examination revealed anomalies typical of AI-generated content, such as unnatural leg movements, unnatural head movements and joined shoes of spectators. AI material detectors confirmed the artificial nature of these videos. Further, digital creators were found posting similar fabricated videos. Thus, these viral videos are conclusively identified as AI-generated and not real depictions of mutated animals.

‍

Claims:

Viral videos show sea creatures with the head of a cow and the head of a Tiger.

Fact Check:

On receiving several videos of bizarre mutated animals, we searched for credible sources that have been covered in the news but found none. We then thoroughly watched the video and found certain anomalies that are generally seen in AI manipulated images.

Taking a cue from this, we checked all the videos in the AI video detection tool named TrueMedia, The detection tool found the audio of the video to be AI-generated. We divided the video into keyframes, the detection found the depicting image to be AI-generated.

In the same way, we investigated the second video. We analyzed the video and then divided the video into keyframes and analyzed it with an AI-Detection tool named True Media.

It was found to be suspicious and so we analyzed the frame of the video.

The detection tool found it to be AI-generated, so we are certain with the fact that the video is AI manipulated. We analyzed the final third video and found it to be suspicious by the detection tool.

The detection tool found the frame of the video to be A.I. manipulated from which it is certain that the video is A.I. manipulated. Hence, the claim made in all the 3 videos is misleading and fake.

Conclusion:

The viral videos claiming to show mutated animals with features like seal's body and cow's head are AI-generated and not real. A thorough investigation by the CyberPeace Research Team found multiple anomalies in AI-generated content and AI-content detectors confirmed the manipulation of A.I. fabrication. Therefore, the claims made in these videos are false.

• Claim: Viral videos show sea creatures with the head of a cow, the head of a Tiger, head of a bull.
• Claimed on: YouTube
• Fact Check: Fake & Misleading

‍

‍

Introduction

With the rise of AI deepfakes and manipulated media, it has become difficult for the average internet user to know what they can trust online. Synthetic media can have serious consequences, from virally spreading election disinformation or medical misinformation to serious consequences like revenge porn and financial fraud. Recently, a Pune man lost ₹43 lakh when he invested money based on a deepfake video of Infosys founder Narayana Murthy. In another case, that of Babydoll Archi, a woman from Assam had her likeness deepfaked by an ex-boyfriend to create revenge porn.

Image or video manipulation used to leave observable traces. Online sources may advise examining the edges of objects in the image, checking for inconsistent patterns, lighting differences, observing the lip movements of the speaker in a video or counting the number of fingers on a person’s hand. Unfortunately, as the technology improves, such folk advice might not always help users identify synthetic and manipulated media.

‍

The Coalition for Content Provenance and Authenticity (C2PA)

‍

One interesting project in the area of trust-building under these circumstances has been the Coalition for Content Provenance and Authenticity (C2PA). Started in 2019 by Adobe and Microsoft, C2PA is a collaboration between major players in AI, social media, journalism, and photography, among others. It set out to create a standard for publishers of digital media to prove the authenticity of digital media and track changes as they occur.

When photos and videos are captured, they generally store metadata like the date and time of capture, the location, the device it was taken on, etc. C2PA developed a standard for sharing and checking the validity of this metadata, and adding additional layers of metadata whenever a new user makes any edits. This creates a digital record of any and all changes made. Additionally, the original media is bundled with this metadata. This makes it easy to verify the source of the image and check if the edits change the meaning or impact of the media. This standard allows different validation software, content publishers and content creation tools to be interoperable in terms of maintaining and displaying proof of authenticity.

‍

‍

‍

The standard is intended to be used on an opt-in basis and can be likened to a nutrition label for digital media. Importantly, it does not limit the creativity of fledgling photo editors or generative AI enthusiasts; it simply provides consumers with more information about the media they come across.

‍

Could C2PA be Useful in an Indian Context?

‍

The World Economic Forum’s Global Risk Report 2024, identifies India as a significant hotspot for misinformation. The recent AI Regulation report by MeitY indicates an interest in tools for watermarking AI-based synthetic content for ease of detecting and tracking harmful outcomes. Perhaps C2PA can be useful in this regard as it takes a holistic approach to tracking media manipulation, even in cases where AI is not the medium.

Currently, 26 India-based organisations like the Times of India or Truefy AI have signed up to the Content Authenticity Initiative (CAI), a community that contributes to the development and adoption of tools and standards like C2PA. However, people are increasingly using social media sites like WhatsApp and Instagram as sources of information, both of which are owned by Meta and have not yet implemented the standard in their products.

India also has low digital literacy rates and low resistance to misinformation. Part of the challenge would be showing people how to read this nutrition label, to empower people to make better decisions online. As such, C2PA is just one part of an online trust-building strategy. It is crucial that education around digital literacy and policy around organisational adoption of the standard are also part of the strategy.

The standard is also not foolproof. Current iterations may still struggle when presented with screenshots of digital media and other non-technical digital manipulation. Linking media to their creator may also put journalists and whistleblowers at risk. Actual use in context will show us more about how to improve future versions of digital provenance tools, though these improvements are not guarantees of a safer internet.

The largest advantage of C2PA adoption would be the democratisation of fact-checking infrastructure. Since media is shared at a significantly faster rate than it can be verified by professionals, putting the verification tools in the hands of people makes the process a lot more scalable. It empowers citizen journalists and leaves a public trail for any media consumer to look into.

‍

Conclusion

‍

From basic colour filters to make a scene more engaging, to removing a crowd from a social media post, to editing together videos of a politician to make it sound like they are singing a song, we are so accustomed to seeing the media we consume be altered in some way. The C2PA is just one way to bring transparency to how media is altered. It is not a one-stop solution, but it is a viable starting point for creating a fairer and democratic internet and increasing trust online. While there are risks to its adoption, it is promising to see that organisations across different sectors are collaborating on this project to be more transparent about the media we consume.

References

‍

• https://c2pa.org/
• https://contentauthenticity.org/
• https://indianexpress.com/article/technology/tech-news-technology/kate-middleton-9-signs-edited-photo-9211799/  
• https://photography.tutsplus.com/articles/fakes-frauds-and-forgeries-how-to-detect-image-manipulation--cms-22230
• https://www.media.mit.edu/projects/detect-fakes/overview/
• https://www.youtube.com/watch?v=qO0WvudbO04&pp=0gcJCbAJAYcqIYzv
• https://www3.weforum.org/docs/WEF_The_Global_Risks_Report_2024.pdf
• https://indianexpress.com/article/technology/tech-news-technology/ai-law-may-not-prescribe-penal-consequences-for-violations-9457780/
• https://thesecretariat.in/article/meity-s-ai-regulation-report-ambitious-but-no-concrete-solutions
• https://www.ndtv.com/lifestyle/assam-what-babydoll-archi-viral-fame-says-about-india-porn-problem-8878689
• https://www.meity.gov.in/static/uploads/2024/02/9f6e99572739a3024c9cdaec53a0a0ef.pdf

‍

Executive Summary:

‍

We have identified a post addressing a scam email that falsely claims to offer a download link for an e-PAN Card. This deceptive email is designed to mislead recipients into disclosing sensitive financial information by impersonating official communication from Income Tax Department authorities. Our report aims to raise awareness about this fraudulent scheme and emphasize the importance of safeguarding personal data against such cyber threats.

‍

Claim:

Scammers are sending fake emails, asking people to download their e-PAN cards. These emails pretend to be from government authorities like the Income Tax Department and contain harmful links that can steal personal information or infect devices with malware.

Fact Check: 

Through our research, we have found that scammers are sending fake emails, posing as the Income Tax Department, to trick users into downloading e-PAN cards from unofficial links. These emails contain malicious links that can lead to phishing attacks or malware infections. Genuine e-PAN services are only available through official platforms such as the Income Tax Department's website (www.incometaxindia.gov.in) and the NSDL/UTIITSL portals. Despite repeated warnings, many individuals still fall victim to such scams. To combat this, the Income Tax Department has a dedicated page for reporting phishing attempts: Report Phishing - Income Tax India. It is crucial for users to stay cautious, verify email authenticity, and avoid clicking on suspicious links to protect their personal information.

‍

‍

Conclusion:   

‍

The emails currently in circulation claiming to provide e-PAN card downloads are fraudulent and should not be trusted. These deceptive messages often impersonate government authorities and contain malicious links that can result in identity theft or financial fraud. Clicking on such links may compromise sensitive personal information, putting individuals at serious risk. To ensure security, users are strongly advised to verify any such communication directly through official government websites and avoid engaging with unverified sources. Additionally, any phishing attempts should be reported to the Income Tax Department and also to the National Cyber Crime Reporting Portal to help prevent the spread of such scams. Staying vigilant and exercising caution when handling unsolicited emails is crucial in safeguarding personal and financial data.

‍

• Claim: Fake emails claim to offer e-PAN card downloads.
• Claimed On: Social Media
• Fact Check:  False and Misleading 

‍

‍

‍