I checked out this app that creates a "podcast" from an article. I put a link to one of my own blog posts there. As I listened to the result, I realized that my face was forming an expression of disgust. Something about it was super creepy. I do not like. Here's the result, followed by the source article for comparison.

elevenreader.io/app/reader/gen

Source: bit-101.com/blog/posts/2024-05

Part of the creepiness was that so much extra stuff was added into what I said. Not just extra words and "facts" but entire concepts I never touched on. "Digital art preservation is in crisis. Artists are losing their work faster than traditional paintings fade." WT actual F??? Apparently my solution "might revolutionize how we preserve digital creativity for future generations". Geez...

It would be one thing if a human were doing this for whatever reasons, but an AI doing it without even knowing why it was doing this, is just weirding me out. It just takes the meaning out of ... meaning. An article is just an algorithmically generated word soup that's formatted correctly enough that humans accept it. This was an eye opener for me on just how messed up AI might be for our culture/society.

@bit101

There's a drum that I've been banging for a long time, that we really needed to normalize things like digital signatures for authentic reporting, where everybody involved would sign the content to confirm that it's legit. It would create a digital chain of custody, in a way, that we could validate.

I even had friends in journalism outright push back against that proposal. They had their reasons, none that I found compelling.

And so we are now getting to the world that I've been fearing for a while, when AI can generate content and we can't really be sure what is and is not human validated. We never put in place the norms that would help us check authenticity, and now we are entering a world authenticity is going to be more and more suspect.

Anyway, yeah posts like these just make me think about how long it's been that I've been seeing this coming and wishing we would take steps to protect ourselves from its implications.

Well here we are.

@volkris Oh yeah. C2PA. My company is heading in that exact direction. We're already working on fraud prevention. Content provenance is a good next step for us.

Follow

@bit101 cool!

I support any effort that promotes that sort of thing.

Although, there is an issue that it has to be societal, not just technological. It's not enough for smart engineers to build the system, we have to have non-technical consumers convinced to look for it and use it.

It just makes me think about how end-to-end encryption is a long-solved problem, technologically, but socially it's still not the standard.

Anyway, here's hoping! But I'm not optimistic, unfortunately.

@volkris Yes, acceptance by the general public takes a long while. My company started with on-line notarization. When I started in 2017, only the state of Virginia had a remote online notarization law. Now most states do and all states recognize it. But the general public is still finding out slowly. It takes years.

Sign in to participate in the conversation
Qoto Mastodon

QOTO: Question Others to Teach Ourselves
An inclusive, Academic Freedom, instance
All cultures welcome.
Hate speech and harassment strictly forbidden.