What are deepfakes and why are they a concern?

23 December 2020, 22:04

Alternative Christmas Message 2020
Alternative Christmas Message 2020. Picture: PA

Channel 4 will raise awareness of deepfakes in its alternative Christmas message for 2020.

Things do not appear as they seem on this year’s Channel 4 alternative Christmas message.

The broadcaster has decided to raise awareness of deepfake technology for 2020, showing a false depiction of the Queen.

So what are deepfakes and why are they of concern?

Alternative Christmas Message 2020
Deepfake of the Queen in Channel 4’s alternative Christmas message (Channel 4/PA)

Deepfakes use ever-improving artificial intelligence and machine learning, allowing people to combine and superimpose existing images and videos of a person to make it look like they have said something they have not.

While these can often be used as humour, they are also open to abuse by those aiming to deceive others, particularly in politics.

Despite renewed concern about misinformation in the past year due to the pandemic, there have been no notable cases of deepfakes used to spread fake news about the coronavirus reported as yet.

It is possible to tell if a video is deepfake as faces do not always look quite right but they are getting harder to spot as the technology evolves.

Facebook came under the spotlight in 2018 for allowing an altered video of US House Speaker Nancy Pelosi to remain on its platform and go viral, which made the congresswoman’s speech appear slurred.

Social network chief Mark Zuckerberg later admitted Facebook took too long to notice the edited video and ultimately decided to crack down on deepfake videos in the lead-up to the 2020 US presidential election.

US House Speaker Nancy Pelosi has been victim of a deepfake video which went viral on social media
US House Speaker Nancy Pelosi has been victim of a deepfake video which went viral on social media (Liam McBurney/PA)

Google revealed in September 2019 that it had created thousands of manipulated deepfake videos with actors, in a bid to help researchers build and train automated detection systems to spot clips.

According to research by Sensity.ai for a Channel 4 Deepfakes documentary into the matter, more than 60,000 deepfake videos were identified on the internet in the last 12 months.

“The UK, it is a very significant target,” said Sensity’s Giorgio Patrini.

“If you look at worldwide statistics, we actually calculated about a 10% of the deep fake video, that we know of on the internet, would feature victims that are British nationals.

“Most of the targets, most of the victims that we see in the fakes are coming from the entertainment industry, the reason why we see that UK up there in the statistic is mainly due to the exposure of the British culture around the world, because that make interesting to feature those people in a video.”

Deepfakes: Can You Believe Your Eyes airs on Channel 4 on December 28.

By Press Association

More Technology News

See more More Technology News

Google homepage

Competition regulator objects to Google’s ad tech practices

A passenger waits for a Tube train at Westminster London Underground station

TfL restricts access to online services due to cyber attack

A purple Currys sign above a store entrance

Currys boosted by AI-curious customers as it takes 50% laptop market share

The Darktrace wesbite

Darktrace chief steps down ahead of £4.3bn private equity takeover

Charlotte Owen

Baroness Owen to introduce law change aimed at criminalising deepfake creation

Hands using computer with artificial intelligence app

UK signs first international treaty on artificial intelligence

The logo of mobile phone network EE is displayed on the screen of a smartphone

EE launches its first standalone 5G network across 15 UK cities

Lord Chancellor Shabana Mahmood signs first legally-binding treaty governing safe use of artificial intelligence.

'We must not let AI shape us': UK to sign first international treaty to safeguard public from risks of artificial intelligence

Visa debit card sitting on a keyboard

Visa unveils initiative to boost consumer protection for bank transfers

A child using a laptop computer

Seven in 10 children exposed to harmful content online – research

Oasis band members Noel Gallagher and Liam Gallagher

Dynamic pricing to be examined by European Commission amid Oasis ticket furore

Amazon's new AI-powered shopping assistant Rufus on a smartphone

Amazon launches AI-powered shopping assistant Rufus in the UK

Gamers play on a PlayStation 4

Sony to take multiplayer game Concord offline two weeks after release

A woman's hnad on a laptop keyboard

Competition watchdog clears Microsoft arrangements with Inflection AI

Health Minister Stephen Donnelly (PA)

Time of expecting social media sites to remove harmful content ‘is over’

An Nvidia sign

Nvidia shares plunge nearly 10% in largest single-day value loss for a US firm