Fact Check: How to spot AI videos by Sora | DW News 

The people who care whether it's AI or not are not the target audience.

This is why as humans we need to talk to each other face to face. Not on a phone, text,or video.

So in the future, every video, voice, and photo we see, we have to think about whether it's AI or not? It's horrible.

I can't think of a single good reason for this to even exist...

Do we need this - I don’t want it

It's horrible. What you see may not be real.

Anti-democratic interests will benefit by making everything questionable. People who make knee jerk decisions with low tolerance for ambiguity will be easily brought to pull their ideological dogsleds, while those who mull and discern will be left out.

One thing I think should be pointed out is that the text in AI-generated imagery tends to be meaningless - case in point, the "hanzi" in the background of the video of the woman walking, not to mention the Japanese and Korean characters arbitrarily mixed in. (Of course the problem is that people unfamiliar with the writing system won't be able to tell)

Why are we even creating AIs to make perfect fakes?

Great, how can i get access to sora? I want to spread some misinformation on the internet.

How do we know YOU are not AI generated?

For some reason I was focused on her hands, waiting for an illogical number of fingers to appear, and at the end of the video she would say "this is an AI video" xDD

these are just the first iterations... they will be flawless before you know it

We are so....

Sora, please redo season 8 of Game of Thrones.

Wonky fingers was the running joke in AI art groups for maybe the first year before Mid Journey and higher end AIs quickly fixed most of that. Sora will correct the hallucinations soon enough and it will be wonderful and terrifying...

Isn't critiquing ai in this specific & public way just giving it more tools to circumvent our detection even easier??🤔

I think we are f...ed.

Excellent work & outreach! SE USA here. My 3 fave sources, no order: DW, UATV, and of course Paul Harrel.

the adice in the first part is useless by the end of this year. and might give false confindence people can see the difference if they look closely. Because soon you won't be able to figure it out like that

But can you detect it if you don't expect AI?

How to detect it for now! Can we detect anything in 6 months or a year?

TikTok will be boring when OpenAI Sora, a US 🇺🇸 product, goes public. It's more exciting when people can generate videos from the text they type. LOL

1:42 "And one more leading NOWHERE, just for show...."

Sora will definitely be weaponised, one way or another and there's no way to stop it. Don't like it, but it is a cost effective way for many companies, to have video clips that won't trigger some random person's copyright claim. It comes down to how hard video sharing platform and government enforcement in how to handle this issue, including the idea of requiring very strict ID check process before a video can be uploaded to be able to pursue bad actors in spreading misinformation

As of 4:06 very interesting, not surprising, and scary. First order of business is that we should all be aware. Maybe the best thing is for everyone to use the internet strictly for two things: business, and and entertainment. After that, then all news, and public announcements should be broadcast or printed in newspapers AND the printed & broadcast content should agree. That might not be a solution, but it's a framework.

Yeah. So this is the 28.8kbps modem. Within 18 months no amount of tech could distinguish AI video from real footage. Hollywood will cease to exist, we will have 8000 new alternative endings to GOT of higher quality than the original series. There will be no such thing as a human "actor" - that is now a dead profession, same as those who refilled the whale oil in the street lamps.

Being able to create my own videos/ short films whatever sure is very tempting. On the other hand yes, there are alot of risks for abuse. Maybe too many to be outweighed by the pros.

staircases that lead nowhere are VERY common in real life

Thanks nerds, you've broken reality.

im wondering if this DW video is not AI generated

Watching this, I kept thinking "jail time".

when technology has gone too far...

The host here was very convincing for an AI model.

When will Sora be available ?

The problem is not the technology, the problem is that people believe it and don't ask if it is true.

I fear people will start to doubt absolutely everything they see online, which will make them passive. How can we act as voters in a democracy when we don't know if the political leaders even exist?

Imagine someone has lost one hand and random kids tell him he's AI, lol the future!

This AI It is the best innovation and tool for filmmakers, and for artists, but it will also have security restrictions,

Technically Ai should authenticate its self on every creation at 100% no miss on Authenticating its self to let us know its Ai with a flasheover digital frames filtered 3d graphics or like one colt or see through clear like 3dmesh and human can use filter to play on as creating to enjoy but know it's Ai creation and all Ai creations should be no miss chance on with all creations as writings music everything it can with authentication as its Ai signature so people know it's Ai and Ai knows we know and we can grow with Ai together. Right :) Gazoo?

Please don't teach your ai that cats can only have 4 paws, and elephants only 1 head.

"See spot run"

Wow... Very interesting topic It's very clear we're living in a new era... The era of A.I 🤖

now everything we see will have to be analyzed by ai to see if its reall

Before asking what we are thinking, give us access to this tool! 🔥🤬

The future is now, old man!

games will look as real as the Matrix

I think it's a good thing. People are going to learn that perception isn't everything and their own senses can lie to them. People are too easily fooled. Also, AI imagery doesn't conform to reality, bodies don't have skeletons, the world can become rather Escher at times. Everything is an approximation based on averages. AI kinda knows the "what", but doesn't know "why". It won't be as accurate as the world until it can model the world, the what, why and how.

"Visually look good"? The horror!

Stop AI. It's like nuclear bomb.

Why are these evil corps even allowed to develop this? Just enforce copyright and make them responsible...

people can olso do insulting video and blame it on unkown person using AI...strict registering with document holding photo same as for trading accounts should be the minimum requirement, just a google account is ridiculus in this matter.

ah yes,. a glitch in the fabric......

Terrifying

A glitch in the matrix

This is creepy, And we told it what it did wrong....

Hasta La Vista !

I am sorry to say, but no other ‘how to’ video is going to age faster. It will take only very few years and it will be possible to distinguish AI content from real one only with complicated software tools.

We decide if we want to be fooled or not ...

😮 on the AI acceptance scale you have reached fear😮 and shock😮 you now understand the AI is so powerful you can't stop it you can't defeat it it can destroy reality and physically basically do anything😮 we need to show you one more thing and I think then we'll enter acceptance😮

The presenter could be AI generated.... I could be a bot....

Madness

so what I really like it watching ai's imagination picture, video, game etc is so great ! you want go see the real go there your self see real life matter place HA

Geez, there's a ton of good stuff this kind of tech can do and there's also a ton of stuff we can do to minimize its harm. For the first one, assisted story-telling, visualized teaching, tailored entertainment, budgeted ads for small businesses. For the second one, practice critical thinking, cross-referencing, skepticism when it comes to provoking news; basically just stop getting depressed/angry and spreading info before checking multiple sources. Also protect the journalists that do honest investigating work for you, stop letting governments bombing them in their homes like the Israeli are doing to the journalists in Gaza.

All the advice here on how to spot out an AI generated video are short sighted and irrelevant, seeming that AI is going to eventually overcome these errors, and once they do, the question is then posed: how do you spot an AI generated video? answer is: you don't

Hhhmmmmmmm…….

Welcome to the matrix.

Interesting

2nd comment !

AI can either save the human race, or. Destroy it.

AI is evil.

These techniques are all useless. Trust no one. Believe in your common sense before acting on any information you see on the internet.

There's a bunch of shorts on RU-vid right now of little cute bunnies that are AI generated and peoole are lapping it up. Mostly women, of course

technology and new inventions always facilitate humans they are not threat for us and they never can replace humans because creator can never be replaced. we should adopt it rater then terfying

I got a good laugh out of your title knowing that most if not all of your tricks & tips for detection were obsolete by the time this video was released. General tip: If your job is done at a desk, your job can and will be replaced by about $5000 worth of hardware per head in the next three to five years. Period. What you see here is the worse it is going to get, it all gets 'better' from here.

Thanks for that. Still a long way to go in this slippery slope to perfection of a technology we could have done without. 👀🌋🌪️🦟