Indian News Agency Sues OpenAI Alleging Copyright Infringement

One of India's largest news agencies, Asian News International, has sued OpenAI in a case that could set a precedent for how AI companies use copyrighted news content in the world's most populous nation. From a report: Asian News International filed a 287-page lawsuit in the Delhi High Court on Monday, alleging the AI company illegally used its content to train its AI models and generated false information attributed to the news agency. The case marks the first time an Indian media organization has taken legal action against OpenAI over copyright claims.

You can't stop the signal

[schwit1]

This seems like a losing battle for copyright holders. Only the high profile AIs will abide by copyright restrictions.

But will they abide by ethics?

[Sebby]

Only the high profile AIs will abide by copyright restrictions.

Even if this happens (doubtful), the issue here isn't just copyright:

alleging the AI company illegally used its content to train its AI models and generated false information attributed to the news agency

So the AI stated falsehoods (ie. "hallucinated"), then said those falsehoods were from the news agency, which never put out those falsehoods.

Defamation much?

Re:

[Local ID10T]

Defamation much?

As a plain understanding -Yes. A false statement was disseminated to a third party, which being attributed to the news agency, could cause harm to their reputation.

As a legal standard -No. Either "actual malice" or "negligence" would need to be established.

"actual malice" means that the defendant made the defamatory statement "with knowledge that it was false or with reckless disregard of whether it was false or not."

"negligence" involves five elements:
-The existence of a legal duty that the defendant owe

Re:

[Sebby]

So how would that work then (not necessarily as defamation) - they did suffer harm, even if not intended. Just like someone causing an accident with another car while trying to avoid an animal on the road - they didn't intend to create that accident, they were simply trying to avoid another type of accident.

(Or perhaps, like involuntary manslaughter - they didn't intend to kill, but still have to be technically responsible.)

Re:

[Sebby]

Well I guess 'negligence' would satisfy the legal standard.

Still wondering what legal standard could be applied to the AI company - it is their product that created these. Just like some company's washing machine exploding - the end result is harm to a third party - someone needs to be responsible, it doesn't just "happen out of thin air".

Re: But will they abide by ethics?

[vbdasc]

Legal standards in India might be different.

Re:

[jvkjvk]

>As a legal standard -No. Either "actual malice" or "negligence" would need to be established.

Actual malice and negligence are quite easily established.

They know it happens. They don't stop it. Therefore it is "actual malice" by your definition.

Negligence:
They have a duty not to defame the paper.
They breach that duty
harm is evident
defendants actions are proximate cause of harm
and cause in fact.

so yeah. legally done.
They know it harms the news agency and still don't stop doing it.

Re:

[ChatHuant]

This seems like a losing battle for copyright holders.

The hacker's (and Slashdot's) rallying cry used to be "Information wants to be free!". This seems to have been turned completely around when talking about using the information for training AI.

Sic transit gloria mundi, innit?