With the ongoing war and the escalating tension across the Middle East, several war videos have surfaced online claiming to be from the actual scene. It is crucial to spot the fake videos as several AI-generated content has been shared online on a day-to-day basis, raising the question of the authenticity of these videos. Today, AI videos are not bad anymore in terms of their realism. In fact, they are almost hard to identify and look very close to real. In this article, we explore five key pointers to spot AI-generated fake videos.
Spot Visual Inconsistencies
It is crucial to spot the inconsistencies in the videos. AI-generated videos could have multiple inconsistencies. For example, the text in the video may have spelling errors, and the characters could appear distorted. Additionally, the body gestures may also look weird and misaligned. Background elements may also change abruptly between the frames. These are tell-tale signs of AI-generated content.
Look For Watermarks
Sometimes, AI-generated watermarks can be burned on the videos. Google’s Veo or OpenAI’s Sora often have their watermarks on in their videos. It is definitely worth double-checking, as some creators are trying to hide and blur these watermarks.
Detect Digital Watermarks
We can spot digital watermarks invisible to the naked eye. SynthID is one of the digital tools to catch digital marks invisible to the naked eye. It is developed by Google and detects watermarks in all content made with Google’s AI tools. Developed by Google DeepMind, the SynthID works on audio, video, texts, and images. Users can upload their media specially images and videos, on Gemini directly and ask if Google AI created it or altered it.
Find The Source
Usually, by image reverse searching to track down where the image or video appeared first online, it is crucial to track platforms that openly admit that they use AI-generated content on social media.
Check With The Experts
You can talk to visual forensic experts, AI developers, and academics to confirm whether a video has been created using the AI tools. Their independent assessment can prevent the spread of unauthentic content.
