The digital twins issue: Gemini and Nudify Apps

Gemini – The Ugly Twin

Google Gemini dropped recently and while much of the world “ooohed” and “ahhed” at the incredible-looking demo, the Internet sleuths quickly dismantled its claims, revealing that it may have been much less sophisticated than it appeared.

In the original video, you get to see a participant chatting with Gemini, talking it through a series of questions while his actions are being observed visually. He draws and it guesses, he plays rock, paper, scissors, and it knows exactly what he is doing based on context. It looked amazing and like a leap forward, approaching a type of AI similar to what we’ve experienced in movies, but without any of the cool dance moves.

The first point of suspicion came from the video description, which included some evidence of manipulation.

Not much later, Google dropped this blog, that gives up the funk.

Aside from the obvious question of “how good is this thing actually?” many people are also asking themselves, “should we be cool with this?”

In reality, all high-quality demos are scripted, polished, edited, and passed through many levels of validation before they ever make their way to the public.

Google was well within the boundaries of that, especially given how open they were about the whole thing.

So why do people feel the ick?

Somehow we’ve acclimated to this post-GPT world very, very rapidly. When we are told to be amazed, the bar has been set higher than ever, and feeling duped just ain’t gonna cut it.

2023 brought us some absolutely world-changing technologies and making something incremental seem revolutionary, is 2022’s gameplan.

But seriously, I was promised robot butlers. Where’s my robot butler?

AI and the loss of privacy, sorta, kinda, but not really

Time released a piece about the surge of apps that take normal photos and use AI “nudify” the people in the pictures.

While people have been editing the heads of famous people onto the bodies of other – more naked people – since magazines (✂️) were a thing, this is somehow a bit harder to swallow.

These are not actually naked pictures of the person in question – or any person in particular. Yet, if someone were to create a realistic portrayal of you, doing something that was not of your character, you’d probably be rightfully upset.

But I don’t think this is the problem a lot of people think it’s going to be.

We’re jaded; the boiling waters have grown to feel tepid and what may feel extreme for a moment will seem normal, very quickly.

This is where things get tricky. We’re coming up on an era where there is almost no evidence that one could produce that couldn’t be explained away by AI.

“We have a picture of you naked” – “No, that was made by AI”

“We have video of you buying illegal substances” – “Nah bruh, that’s AI”

“We have an audio recording of you price-fixing with one of your competitors”  – “AI.”

In other words, plausible deniability is about to enter its heyday.

All of this makes me spiral into the void.

In the meantime, once scammers can figure out how to use AI to make conversational English, a lot of Boomer widowers are in for some trouble.

But hey, with Twitch saying this week, “Why get AI nudes, when you can get the real thing here?” maybe this whole thing will blow over.

That Drip

• Sam Altman is denying that GPT 4.5 is heading our way soon. But I mean, if the board didn’t trust him then why should we?

• Can this guy just go away already?

• Layoffs are happening everywhere, but at Spotify the CFO is taking the heat.

• OpenAI board members shine absolutely no light whatsoever on the Altman coup. None.

• McDonald’s to employ AI to make its food fresher. IS THIS WHY IT COSTS AS MUCH AS IT DOES NOW????

• Microsoft and Anti-Trust Violations – name a more epic duo. I’ll wait.

  

Sponsor Post

Follow Rainwell on LinkedIn!

If you’d like to sponsor future issues: [email protected]