What exactly are AI Doppelgangers, and what are their pros and cons? You’ll get your answers here if that’s what you’re asking.
Have you ever imagined there’s some social profile out there that looks like you, behaves like you, and maybe even talks like you but doesn’t you? Well, that’s possible.
With the rise of generative AI, many things have become possible. Some of these things are beneficial, while others have significant drawbacks. AI Doppelgangers may be one of those drawbacks that need some attention.
Let’s dive deeper, then!
What exactly are AI Doppelgangers?
In the simplest of words, these are digital versions of someone. These could simply be photos or full-fledged talking avatars of you. At times, even full-scale 3D models that may even have the same voice as you can be made!
Anyone with access to basic AI tools can create an AI Doppelganger for anyone they want. Here’s an Elon Musk Doppelganger I created in about five seconds.
The process is extremely easy. At the very basic level, you can simply feed an AI an image and ask it to recreate it. You can even add your prompts to customize the Doppelganger.
Now, I can use any of the dozens of free tools to make the image “talk.” That’s right! Again, it’s as easy as giving the AI a script and clicking “generate.”
Of course, AI Doppelgangers have their pros and cons. That’s what we’ll be discussing throughout this piece.
Benefits of AI Doppelgangers
Let’s start on a positive note. Why would anyone even make a Doppelganger using AI? Well, there are a few obvious benefits.
AI-generated look-alikes save money and time. They save money because you don’t need a camera, clothes, travel/accommodation costs for the person to travel for the shoot, etc.
This means agencies and producers can train AI models instead of hiring the actual people.
Secondly, AI models don’t need to be paid repeatedly. Sure, there’s some production cost with the change in script/scenes, but it’s exponentially less than hiring a person all over again.
Then, AI Doppelgangers do not get tired. You can prompt them 24X7, and the last result is as good as the first one (or maybe even better, as AI models improve themselves with each input).
If your Doppelganger is trained even better, they can interact with your friends, offer advice, and basically “be” you.
The point is, it certainly has its perks.
Problems with AI Dopplegangers
Not everything is rainbow and sunshine with these digital versions of people.
For starters, anyone can create and control anyone. I could create an Elon Musk doppelganger and make him sponsor my project. You could use my images to create my Doppelganger and have an entirely independent social life online.
In simple words, deepfaking someone has become extremely easy. This video and crypto scam surprised the world, underlining how easy this has become.
Of course, there are legal repercussions. But that’s the second problem with AI Doppelgangers. The legality surrounding these are vague. This is primarily because it’s a very new technology (at least for the ordinary person).
Any AI Doppelganger is “your creation” in a legal sense, and it’s generally not 100% identical to whatever reference the creator uses. Couple these arguments with a good lawyer, and you can get away with almost anything (for now).
Another problem is AI’s incredible learning pace. It can improve itself exponentially. This means whatever capabilities it has of creating someone’s AI clone today will get 100 better in the next few months and years.
Simply put, all the benefits of AI Doppelgangers are instantly converted into “problems’ if the counterpart is created without permission.
How can you protect yourself and your data from misuse?
Let’s understand who you need to protect your data from. AI doesn’t hack into your computers to get your information, at least not yet!
So, you first need to protect your data from humans. Other humans may take photos, videos, and other data, feed it into AI, and create your Doppelgangers.
Here are a few tips to help you do that:
- If possible, limit your social media audience. Only a select few should be able to view
- your posts.
- If limiting the audience isn’t possible, include copyright captions. Always mention that your content can’t be used without your permission.
- Manually search for your doppelganger usernames. E.g. if your username is “Alicia2024,” you may search for “Alicia2024backup,” “alicia2024new,” “alicia2024private,” and so on. You may also search for the same #hashtags on social media that you used on your original posts.
The second threat you need to protect yourself from is AI itself. So, while humans feed AI images and data, AI, too, can be restricted from using the media files.
- Use third-party tools to protect your photos and videos. Some tools let you insert a digital signature or mask that may not be visible to the naked eye but makes it hard for AI to use your photo.
- When using AI tools, do not share your input or results publicly.
- Consent forms: Start declining consent for reuse, especially on AI platforms.
Conclusion
The bottom line is that the world isn’t well enough equipped to fend off AI entirely, not just on the Doppelganger spectrum but any other AI feature in general.
Hence, you must be extra cautious for now, at least until we get better regulations, laws, and technology that can help regulate AI so both parties can coexist in harmony.
Do note that AI isn’t bad entirely. It depends on who uses it and what goal is being achieved. On the one hand, AI can help you create content much faster and save money. On the other hand, some third parties may use your photos and videos and create content posing as you without you even knowing!
The point is you can use AI if you want to while taking all the safety measures discussed above and more.
