Iranian campaign planted fake articles in real U.S. publications using ChatGPT
OpenAI said about 100 AI-generated articles were published across at least 20 news outlets as part of an Iranian influence campaign. One opinion piece in the Port St. Joe Star, a tiny Florida weekly, warned of the Iran war under the byline Ervin Hoskins, a name that also appeared on articles critical of the war in larger publications including Daily Kos and Middle East Monitor, which together have more than 3 million Facebook followers.
The bill will follow Americans beyond the fighting itself.
- chvid
"But ChatGPT maker OpenAI said Thursday that Hoskins’s purported work was generated using its chatbot by users based in Iran as part of an “influence campaign.”"
Sock puppet accounts and fake opinion pieces are nothing new.
The new thing is a super centralized "tool" that besides doing your writing for you, logs everything, judges it, and reports to authorities and news outlets as it sees fit.
- OgsyedIE
Given the plurality of self-hosting options there is no reason to believe that it's not that case that every government on Earth is doing it to some extent.
Perhaps the real problem is too much acceptance of shoddy verification.
- xbar
I can think of one media outlet for whom this would be an upgrade to its accuracy.
- thih9
Finally an article that doesn't say "ChatGPT did something" and puts the spotlight on the operators.
Could we keep that wording for other news? If not, could we then say here that ChatGPT planted fake articles, for consistency?
- card_zero
Seems I'm uneducated about how journalism works: you can just write a polished-looking opinion piece and submit it to obscure outlets, who pass it on to news sites and newspapers? Do you get paid? Or if you offer it for free, is that normal, and not suspicious?
- delichon
I prefer an LLM that can steelman any argument. On the whole I think it's more effective to engage with evil than to ignore it. For that purpose these planted articles are a feature. I say that as an enemy of the government of Iran.
- BeetleB
I'm sure there are a lot of differing (valid) opinions on this.
All I can say is that this highlights the need for open weights models.
> OpenAI said it had banned the accounts behind the campaign
If we knew OpenAI would always ban such activity regardless of the actor, I'd be happy. I don't know this.
- tastyfreeze
All news has been a tool of spooks for a very long time. The official CIA Operation Mockingbird started in the late 40s. Now that AI is involved its news? None of it can be trusted.