Synthetic Content (Deepfakes)

Has anyone experimented with deep fakes (artificial intelligence algorithms that create convincing fake images, audio and video)?. 

Advances in synthetic content is on the rise and is raising some big questions about use, ethics and accessibility.

We're moving into an era where seeing is no longer going to be believing

So far, there’s no public evidence of deepfake clips being used to create political disinformation. But a series of stunts have shown what that might look like (just google Trump halloween). While most of us don’t believe everything that we see, a fake video of a politician making threats or doing something disgraceful, if widely believed, could set off an avalanche of hate. Just as dangerous is the possibility that deepfake technology spreads to the point that people are unwilling to trust video or audio evidence.

A new world of verification or a new world of censorship

It's going to be interesting to see how the internet can protect us. And there's some speculation that Blockchain could play a role with a tracking system for the origin of a video, image, or audio clip. The big question is, will companies or governments implement such protections if they might be seen as infringing on free speech? Something that’s a reoccurring conundrum faced by social networking companies when creating policies for extremist content?

The rise of deepfake detectives

According to a new report from Third Way, only 3 in every 1,000 cyber crimes reported in the U.S. result in an arrest. Thankfully, there are actions being taking to minimise crime, Google is making available up to $25 million in grants for nonprofits and researchers with ideas on how artificial intelligence can be harnessed to solve tough societal problems.

On the lighter side of things…

Fellow Fab Lab member, Jonathon Bolitho fed a recurrent neural network with the scripts for every episode of friends and it learned to generate new scenes… 

https://mobile.twitter.com/_pandy/status/689209034143084547

My 2 cents

I propose that because AI never truly lives a human existence any content generated from an AI system will create a feeling of metal discomfort, or the term I prefer cognitive dissonance. AI could highlight our human truths and help us become a better society, valuing human creativity, collaborating with each other and placing lesser trust in machines and the synthetic world. 

"In the field of psychology, cognitive dissonance is the mental discomfort (psychological stress) experienced by a person who simultaneously holds two or more contradictory beliefs, ideas, or values.”

For some solid discussion on AI ethics and superintelligence here is my favourite podcast on the topic. 

https://twimlai.com/twiml-talk-181-anticipating-superintelligence-with-nick-bostrom/

What do you think about Artificial Intelligence and Corporate Social Responsibility? Do you think the makers have a responsibility, or should it be governed first?