The AI dooooooom thread

[crazy canuck]

There is always a risk analysis and balancing that is required.  If lives can be lost in the time that it takes to generate an analysis of an image to determine whether or not it is valid then the prudent thing to do is not take that time, but just shut the system down to ensure safety.

That's why it is a criminal offence, at least in this country, to intentionally make a false report.

But was it a report? I haven't seen anything on how the posts were framed on social media (like did they tag the train company) just reporting on how fake image stopped trains.

That's a good question. How did the information get to a decision-maker with the authority to shut down the system?  I assume it was a report of some sort.

[Tamas]

I know right.
This is being mostly reported as a "AI does bad" story. But I'd even more so read it as a cutting back on employees and ruining resilience story. This bridge was just half a mile from Lancaster Station.  They seriously didn't have someone they could send on a 10 minutes walk to check?

I loved the double layer as well. The article reported on it probably being AI fake because the journalist ran it through an AI tool, relying on its judgement.

I guess eventually the dead Internet will be supplemented by the AI Internet where we watch and read AI reactions to AI content.

And all this will burn our climate down and collapse our economy. Maybe this will be our civilisation's version of the Easter islands statutes.

[garbon]

There is always a risk analysis and balancing that is required.  If lives can be lost in the time that it takes to generate an analysis of an image to determine whether or not it is valid then the prudent thing to do is not take that time, but just shut the system down to ensure safety.

That's why it is a criminal offence, at least in this country, to intentionally make a false report.

But was it a report? I haven't seen anything on how the posts were framed on social media (like did they tag the train company) just reporting on how fake image stopped trains.

That's a good question. How did the information get to a decision-maker with the authority to shut down the system?  I assume it was a report of some sort.

I guess we also can't discount the possibility that someone made and posted the hoax but then someone else reported it to the train company not knowing that it was a hoax.

[crazy canuck]

I guess we also can't discount the possibility that someone made and posted the hoax but then someone else reported it to the train company not knowing that it was a hoax.

That sounds the most probable.  And the most worrisome as people post AI images that become increasingly difficult to detect are "fake".

[Valmy]

So I love Youtube. I have been on it since 2007 or so. I love all the creative and interesting things people make on it but I am starting to get worried. AI slop is really taking over. It is going to fucking ruin it.

And it is frustrating because there are really interesting and creative things people are doing on Youtube with AI but it is just being drowned out by low effort slop. And the slop is only getting more low effort and harder to distinguish from actual videos.

It hasn't entirely happened yet but you can see it coming. A slow motion trainwreck.

Sort of like how bots made the comment sections worthless awhile back. Now the robots are coming for the content creators.

[HisMajestyBOB]

I've stopped buying ebooks or books from unknown authors if they were written after 2023. Too high a risk of just buying slop.

Searching the web also sucks horribly now unless I stick with known sites.

[Josquius]

So I love Youtube. I have been on it since 2007 or so. I love all the creative and interesting things people make on it but I am starting to get worried. AI slop is really taking over. It is going to fucking ruin it.

And it is frustrating because there are really interesting and creative things people are doing on Youtube with AI but it is just being drowned out by low effort slop. And the slop is only getting more low effort and harder to distinguish from actual videos.

It hasn't entirely happened yet but you can see it coming. A slow motion trainwreck.

Sort of like how bots made the comment sections worthless awhile back. Now the robots are coming for the content creators.

Yes.
The bors are definitely there on comments. I uploaded a video a few months back and within seconds of the upload, not long enough to have watched it at all, I had a comment which didn't seem to get what the video about and was ranting about immigrants... Now I had mentioned politicians ranting about immigration a little.... But it wasn't the core point at all. I'd said nothing particularly pro or anti immigration, just anti politicians who don't shut up about it.
I really just don't bother with comments.


Videos.... Yeah. I think it depends on topic. For stuff like warhammer I find tonnes of ai. Bots just reading out fan wiki articles, or even on one occasion a really shit and clearly AI written story that didn't fit in the universe at all.