Privacy and AI

So how does it stand with privacy and AI?

Does everything you enter go straight to America, or worse: to China?

The answer is not black and white. Much depends on which AI tool you use, how you use it and which settings you choose.

First of all: if you do not pay for an AI tool such as ChatGPT, then you pay with your data. Everything you type can be used by the tool to improve itself. That, by the way, is no different from WhatsApp or other free apps we use daily. Most AI tools, such as ChatGPT, process data in the US, but under privacy rules that also apply to European users.

Does your family use AI a lot? Then consider a paid account. You then not only have a more powerful model and better privacy settings, but you can also indicate that answers should fit your (Christian) values. (That last part sometimes requires some adjusting: at one point I got a Bible text with every single prompt, even when I asked how to unclog my toilet…) You can, by the way, turn off the memory function via Settings → Personalization → Memory. Also check regularly what AI has stored about you and your family.

But what I worry about more is a newer development: Sora 2. This is an AI tool with which you can make realistic videos of real people. You upload a photo, and AI makes a video of it in which that person moves and talks. Combine that with voice cloning, and you have a lifelike clip of someone who never said those words. This makes children vulnerable. Think of bullying: classmates can make clips in which your child says or does things that never happened. Be careful with photos and videos of your family that you put online. What is public can be misused more than ever.