[LINK] News Corp hints AI future

[Jan Whitaker]

On 14/08/2023 12:54 pm, David wrote:
> What happens when AutoRantAI (sorry, Tom, that's too good to pass up!) 
> in its value-free trawling through the dark recesses of the Internet 
> publishes some highly defamatory item online, even a fictional one it 
> found on Facebook, or a national secret with an "eyes-only" 
> classification, or proprietary secrets it located through a poorly 
> configured firewall?  Which human carries the can when the case comes 
> to Court?
>
> As I understand the law, it's necessary to demonstrate culpability on 
> the part of the defendant.  And at very least, such issues might raise 
> legal questions bearing on the difference between a publisher and an 
> author.

Publishers. They are the final gatekeeper. That's why the fight over 
Wikileaks exists. Also the arguments over platforms vs publisher when it 
comes to Facebook and others. As long as they don't exercise editorial 
control, they can say they aren't the publisher.

It's messy, though, because the regulators make the platforms exercise a 
degree of content control. It's not like people are standing on the soap 
box in the village square without an intermediary.

As for artificial output with input from supposed actual text sources 
(no matter if the original has nothing to do with the chosen output and 
was contextualised in the original), it's an even bigger mess. BUT it 
would still come down to the publisher -- i.e. News Corp.

In the US (not sure about Australia), corporations earned status as a 
"person". So they can be sued. "Free speech" has its limits.

I am not a lawyer.

Jan