Privacy warriors whip out GDPR after ChatGPT wrongly accuses dad of child murder

HELLFIRE

https://www.theregister.com/2025/03/20/chatgpt_accuses_man_of_murdering/

A Norwegian man was shocked when ChatGPT falsely claimed in a conversation he murdered his two sons and tried to kill a third - mixing in real details about his personal life. …

Regards

px_eliezer_

OpenAI has tried to sidestep its obligations by adding a disclaimer that says the tool “can make mistakes”,

Schmucks.

egeezer

Food for thought; Supposing I set up a legal NIL framework for myself. Would this not provide a legal mechanism whereby I could prevent my NIL from being scraped and used in commercial AI products, data brokers, etc. or receive compensation for unauthorized use?

Somewhat-Reticent

px_eliezer_ the tool “can make mistakes”

And then other tools use those mistakes to make more mistakes, ad infinitum.

px_eliezer_

Somewhat-Reticent

Great fleas have little fleas upon their backs to bite ’em,
And little fleas have lesser fleas, and so ad infinitum.
And the great fleas themselves, in turn, have greater fleas to go on;
While these again have greater still, and greater still, and so on.

https://en.wikipedia.org/wiki/Siphonaptera_(poem)

px_eliezer_

HELLFIRE OpenAI has tried to sidestep its obligations by adding a disclaimer that says the tool “can make mistakes”,

Wow, that sort of disclaimer would be laughed out of court when a doctor is sued for medical malpractice (or when a manufacturer is sued for product liability).

Blackbird

HELLFIRE noyb said OpenAI previously argued it couldn’t correct false data in the model’s output, which is generated on the fly using statistics and an element of randomness.

Events are reality, not statistical imaginings. Facts are specific and real, not random constructs. Using AI to generate “data” in this way as if it were reality is foolish and dangerous.