🤖 SORA AI - Should We Be Worried? 🤖

Artificial intelligence has been a soaring development these recent years. Starting with prompts that can create long, cohesive and almost fully factual paragraphs, to prompts that produce high-quality images -- OpenAI (the creators of Dall-E and ChatGPT) is almost ready to introduce their new service to the public where simple worded prompts can produce hyper realistic videos. This doesn't come out of nowhere obviously as the leaps that OpenAI have taken thus far with their tools and services sort of hint that there is always something bigger in the making.

What is SORA?

Sora is an AI tool built to generate complete videos of up to 1 minute in duration. By simply inserting a text prompt, such as "a house porch with two golden retrievers playing with a tennis ball," theoretically, you should receive a video with content matching the description. OpenAI has revealed a collection of sample videos, showcasing Sora's impressive ability to produce hyper realistic visuals. As you can see from their demo video above, those videos depict mirror reflections, precise fluid motions, and even falling snow particles.

AI-generated video 2 years back 😱

AI-generated video now 😍

Furthermore, the image comparison above clearly shows the immense progress that has been reached by Sora AI -- talking about the jump from a derranged looking Hollywood celebrity eating God knows what, to super adorable puppies playing in snow. It can create videos with accurate lighting, reflections and natural biological characteristics. It is however, by no means perfect. There are still errors such as glitching body parts, elements appear and disappear out of nowhere, things floating when they shouldn't be and others. Though of course, we can argua that there's no stopping them from actually perfecting this tool.

Possible Implications

As interesting and promising as it sounds, there are some aspects that really need to be looked into when adopting this technology. Firstly is how it can be used to spread misinformation. We saw how Deepfake being misused into creating something that spreads false information, so how can we regulate someone from creating a "fox with two heads and nine tails" video and spread it as "this insane new breed shows the end is near". After all, videos are always more convincing than images.

Secondly, it may tamper the organic nature of video-based content social media platforms. How do we stop people from creating multiple AI-generated videos and monetizing it? Video platforms may just lose their charm with the lack of emotions and superficially scripted content. Lastly is how it can be used to create illegal content. Strict monitoring and regulations are absolutely needed in order for people to not misuse this tool to create content that are against moral values.

To summarize...

SORA AI is a promising tool that a lot of people really look forward to. Just as text and image AI generation, it aims to simplify the creation of high-quality meaningful videos. However, just as any other technological advancement, there are threats that may come with its adoption. It shows that there is a clear need (and probably will become urgent soon) for proper AI regulations to prevent users from wrongly abusing the tool. Use AI responsibly. Beep~ bop! 🤖