EN

แชร์

Copied!

verified: “Snow Lotus Grasshopper” clip with 10 million views verified as AI-Generated

25 มี.ค. 6808:05 น.
1
รอบโลก#ข่าวปลอม
verified: “Snow Lotus Grasshopper” clip with 10 million views verified as AI-Generated

บริการเสริมจาก Thai PBS AI

The clip featuring a grasshopper claimed to be a "Snow Lotus Grasshopper" has garnered over 10 million views. In reality, it is just an AI-generated image, and some viewers were misled into believing it was real.

A viral clip of the so-called “Snow Lotus Grasshopper,” with over 10 million views, has been confirmed as AI-generated. Thai PBS Verify found that the Reels video, showing a grasshopper with what appears to be a white lotus flower on its back, was not real. Some viewers were misled into believing it was genuine. 

Source: Reels 

ภาพบันทึกหน้าจอแสดงคลิปที่ถูกสร้างจาก AI

กระบวนการตรวจสอบ

The investigation revealed that a Facebook user named “ครูนิต ชีวิตครูไทย” (“Kru Nit – Thai Teacher Life”), who has 53,000 followers, posted a clip of a grasshopper with what appears to be a lotus flower on its back. The caption read: “Snow Lotus Grasshopper #reels #remix #strangebutbeautifulanimals.” The video shows a white grasshopper with a lotus-like structure landing on a person’s palm. The clip received 320,000 likes, was shared 10,000 times, and has been viewed over 10 million times. 

However, an irregularity in the clip reveals a flaw in the AI-generated content as the person’s hand appears to have six fingers. 

The image was also analyzed using a generative AI detection tool, which indicated a 60% likelihood that it was AI-generated. 

ภาพบันทึกหน้าจอแสดงลักษณะของนิ้วที่เกินจากปกติ

The image was also analyzed using a generative AI detection tool, which indicated a 60% likelihood that it was AI-generated. 

ภาพบันทึกหน้าจอแสดงภาพผลการตรวจสอบด้วยเครื่องมือตรวจจับภาพ พบว่ามีความเป็นไปได้ถึง 60 เปอร์เซ็นต์ที่ภาพจะถูกสร้างด้วย Generative AI

The clip was also reviewed by Asst. Prof. Dr. Rachada Kongkachandra from the Data Science and Innovation program, College of Interdisciplinary Studies, Thammasat University. She noted that the video initially appeared real, as it lacked the usual flaws seen in AI-generated content. However, the extra finger clearly revealed it was AI-generated. (Record link) 

ผศ.ดร.รัชฎา คงคะจันทร์ สาขาวิทยาศาสตร์และนวัตกรรมข้อมูล วิทยาลัยสหวิทยาการ มหาวิทยาลัยธรรมศาสตร์

Determining whether a video is AI-generated may soon require frame-by-frame analysis with greater detail. The rapid advancement of generative AI now allows for highly realistic clips. For viewers unfamiliar with grasshoppers or similar subjects, it’s understandable that they might mistake such content for real footage. 

นายวียะวัฒน์ ใจตรง นักวิชาการประจำกองสัตววิทยา สำนักพิพิธภัณฑ์ธรรมชาติวิทยา องค์การพิพิธภัณฑ์วิทยาศาสตร์แห่งชาติ

In today’s AI-driven era, skepticism is crucial. Don’t believe everything you see, but always verify. 

The image was also reviewed by Mr. Veeyawat Jaithong, a zoologist at the National Science Museum of Thailand, who confirmed that no grasshopper species has a flower or lotus-like structure on its back. The closest natural resemblance in Thailand is the Orchid Mantis, known for its flower-like appearance. 

ภาพบันทึกหน้าจอแสดงภาพตั๊กแตนตำข้าวสีชมพู (ภาพจากอุทยานแห่งชาติเขาสก)
The Orchid Mantis, also known as the pink orchid boxing mantis, is a type of mantis rather than a grasshopper, with body and wing shapes that resemble orchid petals. Despite its name, it doesn’t actually have an orchid on its back. This species can be found throughout Thailand, from the northern to southern regions. (Record link here, here)

What is the impact of this misinformation? 

The video attracted over 10 million views, received more than 320,000 likes, and was shared 10,000 times, with around 3,500 comments. A portion of the viewers genuinely believed the grasshopper was real, highlighting how quickly misinformation can spread and be mistaken for fact. 

ภาพบันทึกหน้าจอแสดงข้อความของผู้ที่เข้ามาแสดงความคิดเห็นส่วนหนึ่งที่หลงเชื่อคลิปปลอม

ข้อแนะนำเมื่อได้ข้อมูลเท็จนี้ ?

Although AI-generated images are becoming more realistic, there are still simple ways to detect them: 

1. Assess the plausibility – Consider if the scene is realistic. AI images often have unnatural pixel patterns, lighting, or depth that differ from real photos.

2.Look for small details – Watch for inconsistencies in text, perspective mismatches between foreground and background, or check the image’s origin using reverse image search, EXIF data, or look for AI service watermarks. 

3.Check for multiple angles – Real events usually come with images from various perspectives. AI-generated content often exists in just one or very few similar-looking images. 

4.Use AI-detection tools – Leverage websites and software designed to detect AI-generated images to verify authenticity. 

Read more AI Fake News & Deepfake examples verified by Thai PBS: 

Fake Clip: “Elephant” chasing a hyena for its ear – https://www.thaipbs.or.th/now/content/1975 

AI Deepfake: Falsely claims a famous doctor was bombed – 

https://www.thaipbs.or.th/now/content/2000 

Fake Image: Giant octopus washed ashore in Indonesia –  

https://www.thaipbs.or.th/now/content/2065 

AI-generated Scene: Falsely shared as real location from the Korean “Squid Game” series– https://www.thaipbs.or.th/now/content/2211 

Learn how to detect AI-generated images: 

How to spot AI images? Understand the pros and cons to build digital literacy –  

https://www.thaipbs.or.th/now/content/2263 

แท็กที่เกี่ยวข้อง

ผู้เขียน