So, how is this different from say other tools (i.e: weapons) used in war, criminal activities and so worth? Do/Are the manufacturers of said products also legally liable for the actions taken by consumers of their products? Or is this somehow different?
In other tools, the manufacturers would notice their weapon autonomously attacking a foreign nation quicker than the three months it took OpenAI to initially alert the authorities about the unintentional attack on Australia's healthcare systems.
This specific manufacturer isn't even aware of the illegal activity it is conducting with its own infrastructure at any given moment, and that's acceptable?
What about the damage it hasn't discovered and/or reported yet? "Trust us! Our technology needs to just grow slower, not halted to be developed under more responsible leadership."
I would also add that less smart weapons of war that don't phone home are more like local models: once you download them then you are probably responsible for its behavior.
The current Big AI situation seems more comical than a simple weapons comparison. It would be like OpenAI having a bunch of missile launchers on their site and giving their customers only a trigger button. And then when someone complains that OpenAI is launching missiles all over the place they just say the missiles are too many and too stealthy to keep track of...
they are both the producer and a user, they have demonstrated gross negligence as a user of the technology, when the monitored people planning school shootings and suicide, while not contacting the authorities.
I would also say a major point of comparison is owning vs renting, eg. if I buy a gun, I own it and posses it - if I use ChatGPT, they still own it, I'm only allowed to use it according to their ToS
It will probably come down to a fitness for purpose vs. intentional misuse argument.
If I shoot someone with a gun, that is my intentional action. The gun didn't do anything it wasn't designed to do, nor expected to do.
If I buy a gun and the manufacturer of the gun did not equip it with standard safety mechanisms, and it kills me or a bystander when I hold it wrong, that's selling something unfit for purpose.
So, how is this different from say other tools (i.e: weapons) used in war, criminal activities and so worth? Do/Are the manufacturers of said products also legally liable for the actions taken by consumers of their products? Or is this somehow different?
In other tools, the manufacturers would notice their weapon autonomously attacking a foreign nation quicker than the three months it took OpenAI to initially alert the authorities about the unintentional attack on Australia's healthcare systems.
This specific manufacturer isn't even aware of the illegal activity it is conducting with its own infrastructure at any given moment, and that's acceptable?
What about the damage it hasn't discovered and/or reported yet? "Trust us! Our technology needs to just grow slower, not halted to be developed under more responsible leadership."
I would also add that less smart weapons of war that don't phone home are more like local models: once you download them then you are probably responsible for its behavior.
The current Big AI situation seems more comical than a simple weapons comparison. It would be like OpenAI having a bunch of missile launchers on their site and giving their customers only a trigger button. And then when someone complains that OpenAI is launching missiles all over the place they just say the missiles are too many and too stealthy to keep track of...
they are both the producer and a user, they have demonstrated gross negligence as a user of the technology, when the monitored people planning school shootings and suicide, while not contacting the authorities.
https://globalnews.ca/news/11676795/tumbler-ridge-school-sho...
I would also say a major point of comparison is owning vs renting, eg. if I buy a gun, I own it and posses it - if I use ChatGPT, they still own it, I'm only allowed to use it according to their ToS
open weights opens a can of worms
It will probably come down to a fitness for purpose vs. intentional misuse argument.
If I shoot someone with a gun, that is my intentional action. The gun didn't do anything it wasn't designed to do, nor expected to do.
If I buy a gun and the manufacturer of the gun did not equip it with standard safety mechanisms, and it kills me or a bystander when I hold it wrong, that's selling something unfit for purpose.
Altman or OpenAI? Weird inaccurate headline, I haven't heard CEOs getting personal fines?
When Apple got hundreds of fines because of illegal behavior, we didn't generally see `Tim Cook`, strange?
I suppose Altman and team forgot to add "make no mistakes" to their prompt eh?