The largest collection of video deepfakes online

Manipulated vision of a child swearing at the First Lady telling her to shut up has gone viral on social media.

The video doesn't appear to use any artificial intelligence and is just an audio dubbing which means it is a cheapfake not a deepfake. The original footage is from a Diwali speech the first lady gave at the White House.

The audio track is edited to include a child inappropriately screaming at a teacher, telling her to “shut the f*** up”  dubbed over the audio of Biden’s remarks. As for the vulgar remarks heard, the original footage is from a 2019 YouTube clip where the child swearing is seen taken out of the classroom for bad behaviour. 

The video was uploaded by a satirical Donald Trump account along with the caption “THANKS, KID!!!” With laughing emojis and has gained nearly 20,000 likes. Comments conclude that the clip is real and condone the child’s appalling behaviour.

The rude remarks, dubbed over unrelated vision, show a high level of disrespect as Biden speaks, however, it is clear the vision has been manipulated, especially at the end as the video is slowed down to make her reaction seem more dramatic. The video quality also deteriorates during the slow motion section.

Cheapfakes are growing increasingly dangerous as they are available to anyone with limited editing skills and technologies, especially in the lead up to US midterm elections.

A young grime performer from Birmingham has been brought ‘back to life’ with deepfake technology in a bid to raise awareness of knife crime after he was murdered.

British 18-year-old, Joshua Ribera was stabbed outside a nightclub in 2003. His mum, Alison Cope now is an anti-knife campaigner. In a bid to make the teaching of youth violence compulsory in schools, she worked with various artists to create a deepfake AI video of ‘Depzman’ rapping about the effects knife-crime has on families.

Depzman toured with grime star Skepta where he built up a huge following in that scene. The high quality deepfake video cleverly uses AI technology to talk about heartbreak, destroyed families and tells kids to think twice before packing a knife. The video, released on SBTV, has spread on social media gaining more than a million views on TikTok and shared by rappers Jaykae, JME and Skepta. It was played on 65 prison TVs around the UK.

Who created the deepfake?

Alison Cope
Jamal Edwards

Was the content disclosed as a deepfake?

Yes

Was the deepfake consensual?

Yes

How was the deepfake created?

Deepfake Software
Video Editing
Voice cloning software
Lip syncing software  

A video that has surfaced on TikTok shows footage of Joe Biden singing the world famous children’s song has been identified as a deepfake.

The PM announces he will sing the national anthem but instead begins to sing the opening lines of Baby Shark in the doctored vision where his lips match the words. A news strap at the bottom of the screen makes it look like a broadcast clip with a logo of an onion in the corner which looks similar to that of the satirical website. The original footage is from Biden’s speech on lowering costs for families at a community college in Irvine, California.

From comments on social media it is clear that some people believe the deepfake was actually Joe Biden singing with some claiming this is ‘proof’ he is suffering from dementia and others baffled how he is still a world leader. The video has been removed from TikTok and Instagram. As new technologies continue to advance, deepfake political videos that spread misinformation and influence elections are an increasing concern. The creator's intentions

Who created the deepfake?

Ali Al-Rikabi
Synthetic Voices

Was the content disclosed as a deepfake?

Yes

Was the deepfake consensual?

No

How was the deepfake created?

GitHub
Video Editing
Wav2Lip
HQVoice
Talknet
Stable diffusion

A marketing deepfake video of Elon Musk has gone viral online, gaining nearly 1,000,000 views in just a few days.

The clip shows Elon held hostage in a warehouse by real-estate investment startup reAlpha and explains the companies mission along with some satirical jokes. The quality of the deepfake isn’t very convincing but the voice sounds fairly accurate.

The video acknowledges that this isn’t the real Elon Musk and that the parody was created as under Fair Use as entertainment and education on short term rentals and fractional real estate model and that no malice is intended. 

Who created the deepfake?

reAplha

Nate Doane
Slack Shack Films

Was the content disclosed as a deepfake?

Yes

Was the deepfake consensual?

No

How was the deepfake created?

Video Editing
Deepfake Software
Video Production Team

A doctored video showing Jill Biden clapping and singing along to an anti Joe Biden chant has circulated online. 

The video is a cheapfake and has been digitally altered to replace the original audio with aggressive boos, profanities and a large crowd of people, swearing, cheering and clapping in unison in disdain for the President. On social media people label Jill a ‘fake doctor’ and say the election was ‘stolen by the dirty democrats and their tactics’.

The original clip, uploaded by the NFL shows the First Lady standing along cancer survivors to lead the ‘Fly Eagles Fly’ first song before the game. Everyone seems to be in good spirits and no profanities can be heard in relation to Joe Biden. 

On closer analysis to the doctored vision, it is clear the clapping and lip syncing is off with the anti-Biden chant. The smiles and hand gestures of each person in view simply doesn't suit the rude chant that can be heard or the pace the audio is being played at. The danger of cheapfakes and shallowfakes lie in the fact that they are very easy to create, unlike deepfakes they use no artificial intelligence and can comprise of recontextualizing, relabelling, slowing and speeding vision down and in this specific case, simply swapping one audio track for a totally different one.  

A political group has created a deepfake ad targeting Mayor Adrian Perkins in north Louisiana in what’s believed to be the first time such technology has been used.

People Over Politics released the anti-Perkins parody ad. In the video, a lookalike ‘student’ who sounds similar and looks similar to Perkins is called to the principal's office after his report card is ‘failing Shreverport’. The principal awards him an F in keeping community safe, creating jobs and laughs when he asks about potential job offers.

The short clip discloses the content is a deepfake and not Perkins. Due to the satirical nature of the content, it is unlikely misleading with comments on social media claiming the commercial was funny. Political deepfakes are a major concern, especially close to election time as voting can be impacted due to digital manipulation.

Ukraine’s intelligence of the Ministry of Defense disrupted Russians’ deepfake attempt to discredit relationships with Turkey.

The Russian agents posed as Prime Minister, Denys Shmyhal of Ukraine, and tried to contact one of the founders of the Baykar defence company, which produces Bayraktar drones, Haluk Bayraktar via a live video call using deepfake technology.

A statement noted it was easy to debunk this case as the perpetrators reportedly had low levels of training and exposed themselves during the conversation by using incorrect Russian pronunciations of familiar Ukrainian phrases. The conversation with Bayraktar didn't go as expected. That's because the Russians were confronted by an intelligence operative who posed as Bayraktar’s press secretary. During the conversation he decided to agitate the imposters by asking them "What happened to the Crimean bridge?"

It is believed the purpose of this deepfake was an attempt to discredit cooperation between Ukraine and Turkey. At the end of the conversation, the Russian operatives were informed that they had been exposed and would be prosecuted.

Who created the deepfake?

Russian agents

Was the content disclosed as a deepfake?

No

Was the deepfake consensual?

No

How was the deepfake created?

Real-time video
Deepfake Software

A breaking news deepfake video has circulated on social media spreading misinformation about Kyrsten Sinema who reportedly spoke on Bipartisanship then shortly after, enthusiastically introduced Donald Trump. 

The clip shows vision of Sinema holding up People magazine with a Fox 10 Phoenix news strap stating “Kyrsten Sinema appears with Donald Trump.” A PaulLeeTeeks watermark can be seen at the bottom right hand corner. The original video is from a 2015 presidential campaign rally in Las Vegas where a Trump supporter can be seen on the stage holding up the copy, not Sinema. 

The creator took to social media describing the deepfake as satirical in nature, addressing ‘newcomers’ who couldn’t grasp the concept of ‘political satire’ with profanities. “It’s fun, it’s funny – and if you’re being honest it’s not that far off from really happening,” Paul Lahticks said. The video is no longer online Lahticks’ Twitter account has been suspended. The response online is mainly negative with praise to NBC News Disinformation reporter, Ben Collins, alerting the public that this content is fake and generating high traction. Several people tweet they will be reporting the content and admitting on first glance it didn't appear fake, but some commended him on his previous parody work.

Who created the deepfake?

Paul Lahticks

Was the content disclosed as a deepfake?

No

Was the deepfake consensual?

No

How was the deepfake created?

Deepfake Software
Video Editing
File Footage

A veteran manga artist was swindled out of more than $500,000 by a scammer who claimed he was Mark Ruffalo.

Chikae Ide was contacted on Facebook by someone claiming to be the Bruce Banner-Hulk Avengers actor, who built an online romance and used deepfake technology for video calls. 

The fake Hollywood scammer asked for thousands of dollars claiming he had money troubles and needed cash for a plane ticket and hospital bills. Ide's daughter intervened to help stop the swindler. Ide reported the 3 1/2-year ordeal that cost her 75 million yen to a nearby police station in June 2021, the culprit is still yet to be found. 

The 74-year-old decided to turn her heartache into a novel and manga, Poison Love, in hopes of helping others realise the dangers of online international romance scams and also to make money to pay her friends back who she borrowed money from for the deepfake con artist.

Who created the deepfake?

Unknown

Was the content disclosed as a deepfake?

No

Was the deepfake consensual?

No

How was the deepfake created?

Real-Time Deepfake Software

A Keanu Reeves deepfake account on TikTok has gained more than 438 million views and six million followers but the account is totally comprised of deepfake videos of the Matrix star.

The videos of fake Keanu show him taking part in trends such as popular dances, odd tips and silly jokes and all clips exceed over one million views with some reaching up to nearly 40 million. Several social media users admit to not truly knowing the validity of the clip with one person saying: “POV: you’re checking the comments to figure out if it’s a look alike guy or some kind of face swap filter.”

Reeves has reportedly not had an active TikTok account since 2020 when it was deleted. The superstar has been quoted telling media that ‘privacy is important to him’. Celebrity lookalikes are very common when it comes to deepfake technology with a major increase when it comes to TikTok trends as we've recently seen with Margot Robbie and Tom Cruise.

‍Who created the deepfake?

@unreal_keanu

Was the content disclosed as a deepfake?

Yes

Was the deepfake consensual?

Unknown

How was the deepfake created?

Deepfake Software
Video Editing

An AI-generated video on the internet has transformed Halle Bailey into a white woman for The Little Mermaid.

Shortly after the trailer for the Disney film was released, a Twitter user claimed the clip had been “fixed” by transforming the black actress into a white person using AI software.

One user acknowledged artificial Intelligence scientist @TenGazillionIQ for ‘fixing’ The Little Mermaid, and turning the 'woke actor into a ginger white girl’. He also boasted that the whole movie could be ‘fixed’ with 4x A6000 in 24 hours. This came alongside claims it was ‘purely for educational purposes’ and not to be misinterpreted in a racist way. But the wider public strongly disagreed. 

Twitter has since removed the tweet from social media after it caused outrage online with one person commenting "so you prefer the CGI uncanny valley. I prefer the woman with talent."

Who created the deepfake?

TenGazillionIQ

Was the content disclosed as a deepfake?

Yes

Was the deepfake consensual?

No

How was the deepfake created?

Artificial Intelligence
Deepfake Software

The latest season of America's Got Talent saw Elvis Presley reborn back to the stage.

In a two-part finale, AI performance duo, Chris Ume and Tom Graham, deepfaked the King of Rock with judge Simon Cowell, who sung "You're The Devil In Disguise" and "Hound Dog" alongside a real-time deepfake with backup singers model, Heidi Klum and actress, Sofia Vergara. Elvis tribute singer, Emilio Santoro, donned an unbuttoned black shirt and golden suit jacket decked out in rhinestones fit for the King of Rock n’ Roll.

Metaphysic took to Instagram to showcase their gratitude. “What an absolute honor to have brought back The King, Elvis Presley, to sing once again,” the duo wrote via Instagram. The co-founders added, “We’re feeling very thankful that we were able to show so many of you the power of our super advanced AI tech tonight, and the tremendous joy it brings.” The YouTube video has gained more than one million views.

Who created the deepfake?

Metaphysic

Was the content disclosed as a deepfake?

Yes

Was the deepfake consensual?

No

How was the deepfake created?

Real-Time Artificial Intelligence
Impersonator
Deepfake Software

‍

A real-time, AI-generated performance of Simon Cowell, Howie Mandel and Terry Crews singing Nessun Dorma was shown on America's Got Talent.

Three professional singers stood in front of cameras which seamlessly swapped their faces out for the judges' faces. 

Chris Ume wants to commercialise deepfake technology and make it readily available to anyone in the Metaverse. Some critics say that popularising deepfakes is dangerous, as they worry bad actors will use technology to spread lies and disinformation. Currently AI technology is minimally regulated, leaving it open to manipulation and exploitation by those with bad intent.

Who created the deepfake?

Metaphysic

Was the content disclosed as a deepfake?

Yes

Was the deepfake consensual?

Yes

How was the deepfake created?

Real-Time Artificial Intelligence
Impersonator
Deepfake Software

Emilia Clarke has been deepfaked into House of the Dragon‘s Rhaenyra Targaryen.

The video face swaps the former Mother of Dragons with actress Milly Alcock and has gained close to 37,000 views and several comments of praise. It gives a sneak preview into what it would look like if Daenaryes Targaryen played the part of her six-generations-off grandmother through artificial technology. Rumours have suggested this may happen.

Who created the deepfake?

Stryder HD

Was the content disclosed as a deepfake?

Yes

Was the deepfake consensual?

Unknown

How was the deepfake created?

Visual Effects
Deepfake Software
Video Editing

A man successfully convinced the world he was an AI creation when he is in fact a human being pretending to be fake.

People have been left baffled by artificial intelligence after a TikTok went viral claiming the user of a CGI-heavy account was actually an elaborate deepfake himself, played by an actor.

In the video, uploaded by Curt Skelton, he poses the question: "Will artificial intelligence replace the role of visual effects artists?" Shortly after he claims “it has,” then he is deepfaked into a woman who explains Skelton is a 'fake character' she made whilst convincingly detailing the procedure which included software such as DALL-E, DeepMotion.com and Deepfake Lab.

The hoax certainly caused chaos online with the video viewed nearly 18 million times and thousands of people commenting about their confusion and disbelief.

A Brazillian TV presenter had her image and voice falsified to spread disinformation in what is believed to be the first 'deepfake' the country has seen as the election campaign period kicks off this year.

The realistic yet doctored video mixes the voice of Renata Vasconcellos reporting from ‘Jornal Nacional’, with results of a false voting poll from old footage.

TV Globo confirmed that Renata’s video is not original and reported it. Ipec, an institute that conducts electoral polls and is cited in the 'deepfake', has denounced the material in the Misinformation Alert System.

The video spread quickly on social media – one clip found on YouTube had more than 1,400 views in just 12 hours and was already confusing voters. “Isn’t this research from 2018?”, asked a user in one of the post’s comments.