free web page hit counter
🛡️
Copyright Notice: This video is officially sourced and embedded from YouTube. For all copyright inquiries, reports, or removals, please contact YouTube's legal team here.
New Scientist

New Scientist

569,000 subscribers

👁 106,281 views

How to spot deepfakes and AI-generated images

Video Overview & Insights

It can sometimes be difficult to spot AI-generated videos such as deepfakes, in which a person's facial expression or speech is generated or altered. Deepfakes' potential to misinform or disrupt democratic processes is huge, especially given we are entering an era where anyone can create them with just a simple text prompt. Andrea Hickerson at the University of Mississippi and Edward J. Delp at Purdue University in Indiana discuss the rising threat from AI-created content and what we can do about it.

AI is a dangerous development that will just get more dangerous an more exploitive

— @GoogleHater-h2o

Learn more ➤ https://newscientist.com/article/2445475-how-to-avoid-being-fooled-by-ai-generated-misinformation

Every video on here uses AI on the thumbnails for clickbait and I refuse to watch them.
I didn’t watch this one-just read the comments.

— @thesnowartistofgreenfield418

Subscribe ➤ https://bit.ly/NSYTSUBS

Assume every video you watch is an AI fake. Until YouTube labels the videos as fake, assume they are fake.

— @quartytypo

Get more from New Scientist:

As a artist it is easy to spot. AI image has a Tone issues they tend to lean heavy on variation of a tonal color. Deep Fakes are Movement issues to many point of subtle movement that is not human.

— @Brooklyn_Hit

Official website: https://bit.ly/NSYTHP

Facebook: https://bit.ly/NSYTFB

I don’t want to kill it, it’s awesome anyway, but I think it’s AI … if AI is as good as that, I don’t mind🎉

— @oksam7667

Twitter: https://bit.ly/NSYTTW

Instagram: https://bit.ly/NSYTINSTA

Spotting deepfakes is becoming an arms race, look for blurry pixels or too good to be true lighting, but eventually, we may only be able to trust content from verified third-party sources

— @TikyTokyFatnSloppy

LinkedIn: https://bit.ly/NSYTLIN

About New Scientist:

There are AI services which do well at detecting AI content. This is much like we need lawyers to protect us from lawyers.
Long before AI or deep fakes, we used to call humans who misled us liars.
Those still exist in multitudes, and prosper well.
The antidote to all this is solid, continuous education instead of lazy, ignorant belief.

— @rabokarabekian409

New Scientist was founded in 1956 for “all those interested in scientific discovery and its social consequences”. Today our website, videos, newsletters, app, podcast and print magazine cover the world’s most important, exciting and entertaining science news as well as asking the big-picture questions about life, the universe, and what it means to be human.

Dennis Nicolai Mimoun Aharouil Jeroen van der Geest

— @MariaMarije-f2n8e

New Scientist

"Pope Fake the First". Great. Who needs a REAL one??

— @Joe_Peroni

https://www.newscientist.com/

More User Perspectives

@

Is there any movement to label social media, like YouTube, with a "watcher beware" rating like movies G,PG, PG+13,R etc.? It would be great to have FC 100% (Fact Checked completely,), CS (complete satire), PB (politically biased), AI 100% and DF (deep fake). Scrollers, elderly, impressionable (young and old alike) deserve to have a monitoring system. Freedom of speech protected while propaganda suppressed. The techie world has more than enough resources to make this feasible. Or is the real world only interested in leading the gullible?

@joahlm762
@

I am noticing that more of the short YouTube videos are A. I. generated and are fake. Sometimes it is difficult to detect the deepfake and A.I. generated images.

@davidgrisez
@

I used Undetectable AI Image Detector earlier today and it actually caught something I missed.

@Shengpretty
@

This channel deserves to grow.

@HitenSanghani
@

Key Bible verses on deception in the last days include Matthew 24:4, which warns, "Take heed that no one deceives you," and 2 Timothy 3:13, stating that "evil people and impostors will flourish. They will deceive others and will themselves be deceived". Other verses like 1 John 4:1 instruct believers to "test every spirit, for many false prophets have gone out into the world," and Revelation 12:9 describes Satan as "the deceiver of the whole world".

@markhovis4
@

Regulators. Mount up!

@trillium.3.24
@

With all due respect to AI technology:
Unbearable irresistible life ahead even for AI technologists and their own children! Strictly my opinion.
There must be some common ground, common sense to live life with harmony, the true human life.
You AI folks are hitech thieves. Sirs and madams.
Oh, one more thing, governments must outlaw this profession.

@vasantsande8644
@

This deserves way more views.

@tien-c4x6p
@

Your personality shines through.

@cainhuba
@

Your personality makes this fun.

@lanhau-e5n
@

I’m glad I clicked on this.

@LygiaBevington
@

Such a feel-good upload.

@LuckAnna-z1y
@

I had so much fun watching this.

@DonaseLinasr
@

This deserves way more attention.

@YuJGrnier
@

You always make the best videos.

@دليكياندليكيان
@

I wanted to "see" exactly what was the "fake" parts. Too much blah blah...use a pointer or something.

@maryannturner6575
@

This video cured my boredom.

@ArielChaysen
@

This made me smile.

@KaylaMoody-c4v
@

This is how you do YouTube right.

@ChauPhuc-w8t
@

This was super helpful, thank you!

@JimenaTorresMondragon
@

Social media has obviously damaged society in a way that it may mean war before we resolve the political damage and toxic emotion. what’s to blame? Greed. Politicians were paid not to regulate.. capitalism has turned into monopoly. it rolled out and mutated within one decade. AI adding onto the stress the world is already under.. I fear truth will disappear. We will no longer know it and greed has no moral. As 5 men get extremely rich the rest of the world becomes more poor. This issue is killing us

@prawnstar9213
@

Criminals always try 😅 to stay one step ahead of law enforcement whom are 10 steps ahead 😂😂😂

@CaptainRLutchman
@

This Technology is dangerous and damaging to society for DEFORMATION

@CaptainRLutchman
@

So much talent in one video.

@JohnsEliga
@

Real Qs how to spot fake ppl ?

@shahedhashemi7912
@

Can’t wait for the next upload!

@NickCharlier-i1g
@

Didn’t think it would work this well, but the Undetectable AI Image Detector is super accurate.

@Shakirakagi
@

The sound quality is great.

@MissimerMonks
@

This video was so relaxing.

@AccurateFPS
@

I can't wait for the next video.

@NormaHarris-i3p
@

I didn’t realize I needed this.

@RanchDolf
@

This was oddly satisfying to watch.

@maximelesagne
@

How did I just find this now?

@McGranaghanAffeldt
@

Exactly.

@RenadHonan
@

I’ll be back for more videos like this.

@GregoryHamond
@

Who’s rewatching this already?

@nhu-w7x
@

I gotta say, that asian grandpa looks very convincing. Good job, AI!

@Vitaliuz
@

Stupid clickbait title. No information on how to spot a fake.

@lt5182
@

So where does it tell you how to spot a dep fake?

@Paul_Holmes
@

There are a couple tips near the start but not a lot.

Words, hands, details, lines lining up, symmetry, wrong color, doesn't make sense, doesn't follow physics, looks too polished or airbrushed or more beautiful than real life. These are things to check that can help you spot them.

With music, the singers aren't quite good enough at sliding from one note to another. Their tone of voice or dynamics of their voice might not match the song or lyrics in parts. The lyrics might not fit quite right, or the music doesn't align. There are also a couple popular voices that folk use on Spotify. Plus, if they use AI art for the album art, it's more likely the songs are AI too. Check how frequently theyve posted songs. Did they ever post before 2025 (maybe also 2024?)? Does the description of their band talk about how awesome and renowned or groundbreaking they are? Does it use general vague statements or reference an award that doesn't exist?

If you look at something, and it feels... off, somehow, it probably is.

@Obrik
@

This story is based on false premise : that we can learn how to spot AI pictures or videos. First off, if tech companies had more than superficial ethics, they would have already included disclosure data in every output. But they don’t. So it’s over to government to regulate the AI industry. My hope is in the EU for this.

@PatHand-og9yd