Deepfakes: 10 Explosive Use Cases ??

What do you think when you hear the word deepfake? If you're like most people the first thing that comes to mind probably isn't business applications.

That isn't strange: 99% of media coverage on deepfakes is still predominantly negative.

Interestingly this is inversely correlated with actual negative use, which is likely less than 0.001% [1].

By far the majority of AI-generated content is used in business, education or simply for fun.

Let's celebrate the real business value synthetic media is generating today.

Here's 8 high-growth applications that are already in use today ??

• AI Voice
• Social Content
• GPT3
• AI Video
• Virtual Influencers
• Synthetic Data
• Visual Effects
• AI Creation Tools
• Code Completion

AI Voice

WellSaid, DeepZen & others are making incredible progress in text-to-speech.

Affordable, high-quality voiceovers is a game changer for business & education.

Millions of minutes generated per month. And listen where we're going... incredible.

Social Content

Reface and Wombat has collectively generated +10b videos.

Each video creating a moment of joy and educating consumers on the synthetic future.

Just the beginning of consumers actively becoming part of the media they consume.

And yes, that's me ??

GPT3

The GPT3 API is *already* powering chatbots, text analysis, idea generators & more.

Copy AI and many other co's are building entire businesses on GPT3.

They help their customers write more and better copy than what they did before.

And they're growing crazy fast.

AI Video:

No camera & film crew required. Just a browser.

SynthesiaIO generates thousands and thousands of videos every day.

Training, helpdesk, corp. comms, marketing...

Each one helping people communicate better than with text.

And we'll be shipping *crazy* stuff... ??

Virtual Influencers

Lil Miquela & friends has been killing it for years. IP worth $$$.

Millions of fans are captivated by her interactive storytelling on IG.

Tooling will make virtual influencers easy to create.

Good storytelling will still be hard.

The Marvel & DC's of Gen-Z...

Synthetic Data

Collecting + labelling real-world data to train neural networks is a real pain (and a ?? TAM).

Synthetically generated data for machine learning democratises access to good data.

Fast growing market especially for areas like face recognition and self-driving cars.

Check this Parallel Domain visualization.

Visual Effects

Chris Ume's work needs no introduction; the creator of deeptomcruise and a bunch of other very cool projects.

Paving the road for a new generation of creators who'll singlehandedly be able to do the work of an entire VFX studio in a fraction of the time.

AI Creation tools

FaceApp, TikTok, Snap & other platforms all use synthetic filters and lenses.

Why? Engagement & virality (look at early Snap) via creative expression.

Just a cool feature or can you build entire platforms on synthetic media?

I'm rooting for Wombat/Reface to pull it off.

AI Code Completion

Github Copilot is built on GPT3. Autocomplete on steroids for developers.

Potential big step towards democratising software development esp. combined with tools like Replit.

How many people do you know that weren't comfortable writing before using Grammarly?

[1]

It's obviously impossible to get precise data.

Harmful use *will* increase and we need to build safeguards. Today it is estimated that there are ~14.000 deep fake porn videos circulating online.