πŸ„ The year of fakes

You can't trust anything you see online after this year

Hey SurfersπŸ„!

Today, let's dive into the AI generated fake world we are waking up to. AI is ready to paint us a beautiful world, but it might make us less happy.

Here’s what’s up with AI:

THE NEWS

🎭 The year of fakes

Last year, Aitana Lopez, an AI-generated Spanish influencer, made waves worldwide. Created by an influencer agency, this virtual girl quickly racked up more than 200,000 followers and secured lucrative brand partnerships, bringing in about 10,000 euros a month for its creators.

But Aitana was just the beginning. Nowadays, crafting photo-realistic portraits with a consistent appearance is a piece of cake after a bit of study. A Reddit user recently showed how easy it is to create fake verification pictures, including hand-drawn letters on the body or holding an ID.

We can say for certain that image generation got to a point where it’s hard or even impossible to tell them apart from real ones.

But videos are also getting easier to fake. A researcher demonstrated this by uploading a 30-second clip of himself to HeyGen, then letting the AI create a new video with completely fabricated content spoken by him. Impressively, the AI also translated his speech into Italian and Hindi, maintaining his voice.

And if you think live calls and videos are safe, think again. I talked about this app in yesterday’s newsletter. Pretend+ shows how simple it is to swap faces in real-time. While it's not perfect yet – the hairline sometimes flickers, and complex backgrounds can be tricky – it's likely just a matter of time before these issues are ironed out.

Where does this leave us?

We might be entering a digital dark age, where we can’t trust what we see online at all. This year fake images and videos will have a huge impact as the US, India and many other countries face elections.

And society does not have an answer to this new phenomenon yet. The erroding trust in online platforms and communities might just make the bubbles we live in even stronger.

As for our body image, the impact isn't looking too good either. Think about it: influencers, online models, OnlyFans creators - they've been upping their game with filters and all to set some pretty unreal beauty standards. But now, with the arrival of AI-generated supermodels, the beauty bar is being pushed even higher.

This girl is AI, but has 180k followers and gets DMs from celebrities. (AI face on a real video)

And as they pop up in our feeds on Instagram and Tiktok, we’ll be comparing us to them as well.

That probably means we’ll be using filters and stuff even more, making our online look drift even further from our real-life look.

This might lead to people, especially those who feel unattractive or shy, preferring phone communication with filters over real-life meetings, and opting for virtual intimacy over physical encounters. And at that point chatting with an AI might not feel that different from talking to a real person.

Alright, things took a bit of a gloomy turn there, didn't they? But here's the deal: this whole digital revolution thing is happening, whether we're ready for it or not. It's high time we start talking about what this all means – like, how do we hang onto what's real when there's a flood of fakes out there?

These fakes aren't just playing tricks on our eyes; they're gonna twist our ideas about beauty, mess with our body image, and even skew how we see the world and politics. It's a lot to take in, but we've gotta start tackling it.

ONE MORE THING

OpenAI claps back at the New York Times

OpenAI came out with a blog post in self-defense after the New York Times sued it over the Christmas holidays.

Quick summary:

  • Training is fair use, but we provide an opt-out

  • "Regurgitation" is a rare bug we're driving to zero

  • The New York Times is not telling the full story

  • Like any single source, their content didn't meaningfully contribute to the training of our existing models and also wouldn't be sufficiently impactful for future training.

⌚ If You Have One More Minute

AI Art of the Day 🎨

Magnific can now upscale any image to 10,000Γ—10,000 pixels, perfect for large-scale prints. You can see more examples by clicking on the tweet.

πŸ„πŸŒŠπŸ„πŸŒŠπŸ„πŸŒŠπŸ„πŸŒŠπŸ„πŸŒŠπŸ„πŸŒŠπŸ„πŸŒŠπŸ„πŸŒŠπŸŒŠπŸ„πŸŒŠπŸ„πŸŒŠπŸ„

That's all for today, folks!

If you enjoyed this, please consider sharing this hand-crafted newsletter with a friend.