Not diminishing Miller’s destruction but an AI powered breaking news service could absolutely be more destructive than she ever was, if enough people were relying on it.
Why do we trust human journalists exactly? Very few of them are interested in truth. This article is motivated by the same existential anger we see from artists. Human journalists have a horrible track record spreading war and misery and lies. Plus sounds like this guy would be the type to cheap out and use GPT-3.5 and save money on tokens. We're just going to write off journalism done by GPT-7?
Well, by default we don’t. We typically trust human journalists either because they produce evidence alongside their work or because they have a good enough track record of being right that we believe they are this time too. It’s a pretty fuzzy thing.
To me the core failure in the concept of AI news is that all AI does is regurgitate things other people have written. I know, I know, “that’s all journalists do lol” but in reality journalists do a lot of boots on the ground reporting, they conduct interviews, so on.
Even if GPT-7 totally solves problems like hallucinations there’s still the fundamental problem of it having no first hand knowledge. And very specifically in the world of breaking news (which human journalists get wrong often too) it feels immensely risky.
None of those problems are solved by using humans instead of AI. Very few journalists are doing "boots on the ground" interviews. They're lazy. And why can't an AI do a zoom interview?
> Very few journalists are doing "boots on the ground" interviews. They're lazy.
Sigh.
Of course there are a lot of lazy journalists. There are a lot of lazy everything. There are also journalists putting themselves in significant danger reporting in warzones. Is all of their work invalid because some lazy people write clickbait? No. The answer is no.
> And why can’t an AI do a zoom interview
Not everyone is available on Zoom. Not everyone wants to be interviewed. Some interview subjects have to be tracked down and forced to answer questions.
In a breaking news scenario it’s incredibly valuable to have trusted physical presence. Imagine the next 9/11. How will we be able to trust AI isn’t being fooled by deepfake videos posted to YouTube?
> MALONE: Yeah. We had told the AI, write five interview questions that would help an audience understand this academic paper. And we were going to just read those questions word for word to these academics. And Jeff and I were a little nervous, to be honest, 'cause would they figure this out, and if they did, were they going to be mad at us?
> ...
> GUO: Yeah. So Dan starts talking to us about AI and the kinds of jobs it might replace, but he goes out of his way to reassure us. He says, you two, Kenny and Jeff. AI probably won't be coming for your jobs.
> GROSS: That AI won't be able to ask a question as incisive as the one you just came up with. So perhaps you're safe.
> MALONE: Hey, hold. On, I got to - we were going to - I have to jump in. Are you aware of what's happening? Because AI 100% generated all the questions that we asked you.
Have you considered the possibility that the people you are interviewing may be lying? If the CEO of Dow Chemical gets on Zoom and says the things they’re dumping in the river are safe, what is your AI journalist going to do?
yeah they do the same thing with humans. except human journalists give you a false sense of security if it's parroted by NYTimes' Judith Miller (or 1000s of other examples)
on the other hand, that's what the pulitzer prize for public service is for, to highlight the efforts of journalists who do precisely this; the truth, and the willingness and effort to put their careers, and sometimes lives, on the line, for the truth - or so my idealist side would like to believe.
Humans are (not often enough in practice, but in principle) accountable. Look at Fox News and Alex Jones paying out over $1 billion for their lies about the election and Sandy Hook, respectively.
I don't think anyone is alleging the majority of BNN was not "fine." It's some tiny percentage that was apparently controversial - just like human journalists.
Judith Miller is a good example to look at if you're having trouble with this
