How Adobe Is Fighting FAKE Images

C2PA explained: We delve into the intricacies of the C2PA initiative and how it aims to combat the proliferation of fake content in the digital world.

AI’s ability to create convincing fake content is blurring the line between real and make-believe.

From misleading videos to alarming photos like a recent Pentagon explosion hoax that led stock markets to crash, it was, in fact, just an AI-generated image.

And there were thousands of people who were fooled by this rather stylish image of the Pope in a puffer jacket.

Fighting Back with C2PA

But fear not.

A coalition called C2PA with tech giants like Google, Adobe, and Microsoft is fighting back.

Join me as we explore how they’re shaping the digital world and how they’re protecting truth on the internet.

If you’re new here, I’m Samson Vowles.

And on this channel, we talk about AI and the future of technology.

The Importance of Detecting Fake Content

Recently, during the Turkish democratic elections, a video of the opposition leader was released, purporting him to be partaking in some sort of sexual act.

This undoubtedly had a huge effect on the opinions of the voters.

But who can prove whether it was real or AI-generated?

This is the importance of this new technology.

The Proliferation of AI-Generated Content

So let’s take a look at the problem of fake content on the internet.

According to a concerning report from Europol, an alarming 90% of internet content could be crafted by AI by 2026.

Already, AI-fueled spamming news sites are sprouting like wildfire, mirroring real-life occurrences where manipulated imagery has misled people.

And so the question arises.

How do we combat this hyper-realistic misinformation?

The Solution: Digital Media Nutrition Labels

Well, for this solution, I want to imagine the nutrition label on your favorite piece of food.

Just as nutrition labels on food packages inform consumers about the ingredients and nutritional value, C2PA provides a label for digital media.

This will detail its origin, its creation, and modifications.

It’s a proven way for consumers to understand what’s inside the content they consume, whether it’s been generated by AI or whether it’s a truly authentic piece of content.

Some brilliant minds came up with this novel solution by using tools like metadata, watermarks, and clever technical tricks.

We are going to be able to separate the fake from the real, or at least this is what they are claiming.

Google’s Role in Verification

Google’s role is stepping up, planning to attach written disclosures akin to copyright notices beneath AI-generated images.

Another tool is the About This Image feature, allowing users to explore the history of when and where it first appeared and where else it’s been used online.

As well as, seeing what changes have been made to it.

Adobe’s Innovation

Next up is Adobe’s innovation.

Adobe’s Generative Fill, a smart AI-powered feature in Photoshop, uses content credentials to ensure changes in an image are accurately recorded.

And this is where the clever bit comes in.

It will automatically stick it to the file wherever it goes.

Now Open AI, the company behind Dall-E and ChatGPT, has created a watermark that will be left on all of its creations, adding this extra layer of transparency.

C2PA Explained

Let’s take a look at how C2PA actually works.

You’ll be able to hover over an icon that will be embedded into the quality of the image.

This will then reveal the history of the image on how it’s been adapted over time.

You’ll easily be able to see when it was created and what changes have been made.

Specifically, if it was AI-generated or it’s been adapted by using computer programs.

Now this will also be used across video content for checking the validity of news articles and news videos.

This is particularly important as deep fakes appear of national leaders.

And this recent report stated that generative AI poses a real threat to the stability of worldwide democracy.

However, this is not a foolproof solution yet.

Cunning minds could tamper with watermarks and metadata.

Not every AI system might be so transparent.

As we’ve seen in the rise of online conspiracy theories, people often disregard facts for forces.

But AI disclosures presented as more neutral might guide us through fake versus real content online.

The Coalition of C2PA

Members of the coalition of C2PA, including Microsoft, Adobe, the BBC, and Intel, have clearly stated that their mission is to create an open standard.

This will unite all of the efforts to prove provenance together and allow everyone to recognize the true origin of different media on the websites that we visit.

The Future of C2PA

One last challenge is that companies will have to adopt the C2PA standard for this perfect Utopia to unfold.

High-tech giants like Adobe, Microsoft, and others are already aboard the C2PA train.

Their unified vision is to create an icon signaling verified content.

You’ve got to swipe right on a dating app headshot to make sure there’s been no photoshopping of those.

Or if you’re looking at a breaking news photo on Twitter that seems extremely shocking, you’ll be able to prove it instantly.

But you might be thinking about what’s going to be next.

Well, this information is focusing on images and videos.

We will start to see this also take on text environments.

The impact of C2PA could be significant, offering a beacon of authenticity in our vast digital landscapes of distrust.

If you liked this article, ‘C2PA Explained,’ check out the articles below.

How to Thrive in the AI Revolution (Adapt or Die)

Advanced Prompt Engineering: 4 Mindset Shifts

AI Podcast Automation With Riverside: Saves Me 100+ Hours a Week