TikTok just dropped a new realistic beauty filter

All of those things have had extremely perverse incentives leading people to be dishonest for a long time, long before the internet in the case of photo/video (see Loch Ness Monster, Sasquatch video). I don't see how AI tools have much impact. Most people knew this about AOL chatrooms in 1996. Does anyone actually think internet interactions are mostly authentic, or are you speaking on behalf of some imaginary clueless person?

On IRC/AOL, you don’t see a face. But still, probably some real person did write what you read.

Until a few years ago, if I looked at the FB/IG/anything from a random person, I would assume most photos to be real, and most text to be really written by them. You could edit some photos and have some content ghost written, not everything.

Now a random person can fake everything. Would you still follow that “person”?

Previously specifically real-time video was hard to forge, that tech was out of reach of a typical person who could not hire actors and do seamless real-time video composition.

Now thus has changed. The person you have just interacted with in a video call, with authentic human reactions, etc may as well be completely a visual fake, and have spent $0 to achieve that.

I suppose digital photos already have hard time to be admitted in court:i think most digital video coverage will soon also be inadmissible. I've heard of places where high-stakes security photo registration has switched back to chemical photography, exactly to make it demonstrably harder to forge, given that the footage us physically well-protected.

All around us gets more and more virtual, no matter whether we think, as consumers, it should be, or not. We'll have to deal with it.

> I would assume most photos to be real, and most text to be really written by them. You could edit some photos and have some content ghost written, not everything.

Why does it matter to you if the text was written by them, or by a ghost-writer, or if they just regurgitated whatever their sponsor of the week wants them to say?

As soon as a profit motive's involved, you can't actually trust a media personality. It doesn't matter one whit to me whether it's an actual person shilling from a script, or if its an artificial person shilling from that same script.

> Previously specifically real-time video was hard to forge

But paying someone to say what I want you to hear, into a camera was, and still is, very, very easy.

The incentive was always there but now the cost is lower.

There have always been models and airbrushed photos, but most people quickly learn that Emily Ratajkowski is not going to return their texts and move on to real live human beings in their own cohort.

Now you make every average person into a model but only on TikTok, what happens? Probably more time spent on TikTok instead of face to face interactions with local people, at a minimum.

What happens to teenage girls when they see not just models but all their classmates rendered pretty by the machine and then compare their own reflection in the bathroom mirror?

Making deleterious things cheaper isn't great.

Money is difficulty too. If fake video is ten dollars an hour, or one dollar an hour, that's a lot easier in practical terms than hiring an actor with high availability and a willingness to scam people.

If it's just speaking a preset text, yes. Doing a dialog in a way you specifically want, without a pre-made plot, is rather hard or impossible with an actor. Now you can put your brain behind a made-up face, in real time, for peanuts.