Extremely realistic videos that show someone doing and/or saying something that they never did nor said. For example, celebrities’ faces are being inserted into porn photos and videos while recently, a deep fake video of the actor Tom Cruise playing a guitar and singing went viral. He never did this. But it appears to be him. Somebody else created this video. It went viral. We have just had our first trusted TV News Presenter make a video about a Keep Fit Product. She didn’t, but someone else used her image, body movements, lip synch and tone of voice.
They say that ‘Tech is neutral’. So, like marketing, you can use it for good or for evil. You have good actors and bad actors. Here are some potential advantages and disadvantages of deepfake.
Advantages of Deep Fake
- Enables global advertisements to be tailored and ‘dubbed perfectly’ to local languages and local brand names (see Just Eat and Snoop Dog).
- Improve internal comms by enabling CEOs to speak to their staff around the world in perfect local language.
- Act in a movie after you are deceased
- Direct a movie using a younger version of you or a younger version of another actor.
Disadvantages & Dangers of Deep Fake
- Catastrophic damage to public trust and to markets could come from deep fake attacks.
- Cynical audiences will neither trust nor believe in any news or evidence on camera or video.
- Non-cynical, ignorant audiences will gladly believe deep fake videos that reinforce their opinions and values.
- International opponents will use synthetic media to sow the seeds of distrust plus reinforce anti-democratic values.
- Terrorized women have had their faces faked and inserted into pornographic images and films.
- Deep fake attacks – on markets and trust in markets
- If everything can be deep-faked. Then everything can be denied. Worse still, untrue accusations can have ‘proof’ (deep-fake proof).
- Hijack your bio-metrics
- Boost fraud – intimidate/blackmail celebrities with fake porn or any obscene imagery
- Destabilise trust
Do the Disadvantages Outweigh the Advantages?
What do you think? Can deep fake be used for good only? Have a look at another chilling application of data, untruths and destruction of society in The Dark Arts of Marketing.
‘Just Eat’ Deep Faked their Australian Ad
Just Eat (Snoop Dog) ad was changed to its Australian subsidiary brand name – Menulog. They seamlessly switched out the word ‘Just Eat’ for ‘Menulog’ so that Snoop Dog appears to say ‘Menulog’.
Are Governments Ready For Deep Fake Attacks?
No they are not. US Military, Law Enforcement and intelligence agencies have been monitoring Deep Fakes for years. During a 2019 hearing Senator Ben Sasse asked ‘if the USA is ready for the onslaught of disinformation, fakery and fraud?’ He continued: ‘Are we organised in a way that we could possibly respond fast enough?’ The Director of National Intelligence (DNI) in the USA, Dan Coates acknowledges that it poses a major threat to the United States and ‘it is something the intelligence community needs to be restructured to address.’
How do they create deep fakes?
They filmed a Tom Cruise impersonator who also looked very like him. They only deep-faked Cruise’s face. They digitally stitched Tom Cruise’s face onto the face of the impersonator on video. They start with a library of images of Tom Cruise – every possible angle and include all the expressions and all the emotions of Tom Cruise. They then train the machine with this data set.
Training means it’s going to analyse all the images of Tom Cruise. All his expressions. They then compare these to the impersonator. The computer then teaches itself – so that when the impersonator is smiling, they recreate Tom Cruise smiling. They use AI & Deep Learning to create fake content or ‘deep fakes’.
GANs General Adversorial Networks
It is like setting up a game with two AIs competing to produce the best fake synthetic content – as these two new combat each other – one tries to generate the best image – the other tries to identify or detect where it could be improved – continually improving until you cannot tell the difference!
What have these got in common
(apart from being all white)?
Answer – they are not human. They do not exist. A computer created them. They all come from ThisPersonDoesNotExist.com . Creating characters, avatars, bots and robots is already happening. Some bots are created others are partly regenerated (or reincarnated) from the deceased as demonstrated in Project December as well as Microsoft’s patent for digitally reincarnating dead people as chat bots (if the families give permission) as highlighted in the Chinese Girlbot with 465m boyfriends.
Deep fake acceleration
Cadence, tenor and syncopation are now being incorporated into fakes. You just type the words into a computer and the words appear to come out of someone else’s mouth! Estimates suggest that within 5-7 years, the average youtuber and tik-tok-er – will be able to do what elite top hollwood studios and advanced AI companies can do now.
Synthesia Avatars a UK company appears to be able to do this as they replace cameras with code. And there are many others…coming.
What must you do…..?
- Watch this excellent, albeit frightening, Bill Whitaker report from the 60 Minutes series (the video is only 13 minutes).
- Understand what a deep fake is and what is synthetic media?
- Discuss, how do we inoculate ourselves & understand that this content is coming and exists – without being cynical – without losing trust in authentic media?
- Explore/discuss how do we authenticate stories, photos, videos, media, journalists, bloggers, tweeters ….?
- Read: Nina Schick Deep Fakes – the coming infocalypse.
- Understand how AI can create dialogues and even relationships between humans and avatars – read Chinese Girlbot with 465m boyfriends.
Bill Whitaker reports. “60 Minutes” is the most successful television broadcast in history. Offering hard-hitting investigative reports, interviews, feature segments and profiles of people in the news, the broadcast began in 1968 and is still a hit, over 50 seasons later, regularly making Nielsen’s Top 10.
Subscribe to the “60 Minutes” YouTube channel: http://bit.ly/1S7CLRu
Please do post a comment and I will reply to you.
If you liked this, you might also enjoy:
Facebook Data: How it was used by Cambridge Analytica
Chinese Girlbot with 465m boyfriends
If you would like me to send you any of my future posts by email – please enter your email address in the ‘Subcribe and Get Cutting-Edge Alerts’ section (top right corner).
Join my free 30 minute Friday chat about AI, Innovation and Ethics in Marketing and Business.
1pm (Dublin/London time) on my Linkedin Profile (scroll down to Activity and then select ‘Events’) .
I also issue a 60 Second Summary of all previous chats here.
Perhaps also connect with me on Twitter Linkedin Instagram Youtube .
Consider becoming a SOSTAC® Certified Planner? Add a new skill set. Stand out from the crowd.
Great article! This is scary stuff. The negatives far outweigh the positives in my opinion. We’re already seeing the dangers and how people are influenced by this, mostly negatively. I think we’re facing a significant threat to democracy and law and order. What happens in the US in the next few years could significantly change the world as we know it….and that’s just the tip of the iceberg.
I agree Brian. For me, China is investing, training and embracing AI totally and, in my opinion, is already surging ahead. However, the conundrum, is that as other countries compete in AI, AI capabilities will rise higher to a point where AI outsmarts humans and ‘Unintended Consequences’ may occur.