@carnage4life This further reinforces the real target effect of disinformation as tool of propaganda - it’s not to make one believe in something, but to make one believe nothing.

@carnage4life I recall one of such detectors concluded that the bible is ChatGPT-generated (and with no less than 100% certainty) 😆

@carnage4life

Ah. Tech bros assuming they can do the work of journalists simply by looking at a picture.

Who needs context, sources, intuition and knowledge.

If you look at the pixels hard enough, some sort of truth should eventually emerge.

@carnage4life@mas.to to detect if something is ai-generated or not, flip a coin. according to my dice roll, its probably more accurate

@carnage4life Second layer #disinformation is definitely a big concern, there's a weird cognitive dissonance that enables people to be easier fooled by disinformation posing as something countering disinformation.

Another potential that gets dismissed is the second order impact of purposefully weak disinfo.

Outrage at weak arguments or info can create a flood of people commenting with counter arguments, the impact effectively acting like a ddos amplification attack. People go with the flow.

@carnage4life This is part of Steve Bannon's "flood the zone with shit" strategy... Not necessarily to make people believe the lie (though that's a bonus), but to make people distrust the truth and give up caring.

@carnage4life another reason to dispel the efficacy of "AI detectors".

@carnage4life ten 90% accurate fake detectors and a motivated reasoner...

edit: also it's a black box. The machine learning could be investigated, probably,(tho difficult..) but not if it's running on someone-elses server..