Blog Post #6: A discussion on the advances of AI video creation using Sora
If you have ever tried to create a video using AI, or even seen a video that was created using AI, you know that they can sometimes come out looking a bit unusual. This can be caused by a few factors. The first factor that can make an AI video look unrealistic is a lack of object permanence, meaning that certain features in a video might disappear after a few seconds when they’re not supposed to. Another reason why AI generated videos look weird is because of a lack of “flow” in the video. This means that objects in the video move in a way that they wouldn’t in real life, or even that they just move in a choppy way. OpenAI has just released a new text-to-video model that hopes to combat these issues. The name of this model is Sora, and it is now being known for exceeding what we know the skillset of AI to be. This AI is far better than others that can generate videos due to its sheer computing power, increased resolution, and ability to create videos in a variety of aspect ratios and frames. It is now being debated whether or not this advancement in technology will do more harm than good.
This post was proofread by ChatGPT.