Journalists at 'The Atlantic' Demand Assurances Their Jobs Will Be Protected From OpenAI (msn.com) 57
"As media bosses scramble to decide if and how they should partner with AI companies, workers are increasingly concerned that the technology could imperil their jobs or degrade their work..." reports the Washington Post.
The latest example? "Two months after the Atlantic reached a licensing deal with OpenAI, staffers at the storied magazine are demanding the company ensure their jobs and work are protected." (Nearly 60 journalists have now signed a letter demanding the company "stop prioritizing its bottom line and champion the Atlantic's journalism.") The unionized staffers want the Atlantic bosses to include AI protections in the union contract, which the two sides have been negotiating since 2022. "Our editorial leaders say that The Atlantic is a magazine made by humans, for humans," the letter says. "We could not agree more..."
The Atlantic's new deal with OpenAI grants the tech firm access to the magazine's archives to train its AI tools. While the Atlantic in return will have special access to experiment with these AI tools, the magazine says it is not using AI to create journalism. But some journalists and media observers have raised concerns about whether AI tools are accurately and fairly manipulating the human-written text they work with. The Atlantic staffers' letter noted a pattern by ChatGPT of generating gibberish web addresses instead of the links intended to attribute the reporting it has borrowed, as well as sending readers to sites that have summarized Atlantic stories rather than the original work...
Atlantic spokeswoman Anna Bross said company leaders "agree with the general principles" expressed by the union. For that reason, she said, they recently proposed a commitment to not to use AI to publish content "without human review and editorial oversight." Representatives from the Atlantic Union bargaining committee told The Washington Post that "the fact remains that the company has flatly refused to commit to not replacing employees with AI."
The article also notes that last month the union representing Lifehacker, Mashable and PCMag journalists "ratified a contract that protects union members from being laid off because AI has impacted their roles and requires the company to discuss any such plans to implement AI tools ahead of time."
The latest example? "Two months after the Atlantic reached a licensing deal with OpenAI, staffers at the storied magazine are demanding the company ensure their jobs and work are protected." (Nearly 60 journalists have now signed a letter demanding the company "stop prioritizing its bottom line and champion the Atlantic's journalism.") The unionized staffers want the Atlantic bosses to include AI protections in the union contract, which the two sides have been negotiating since 2022. "Our editorial leaders say that The Atlantic is a magazine made by humans, for humans," the letter says. "We could not agree more..."
The Atlantic's new deal with OpenAI grants the tech firm access to the magazine's archives to train its AI tools. While the Atlantic in return will have special access to experiment with these AI tools, the magazine says it is not using AI to create journalism. But some journalists and media observers have raised concerns about whether AI tools are accurately and fairly manipulating the human-written text they work with. The Atlantic staffers' letter noted a pattern by ChatGPT of generating gibberish web addresses instead of the links intended to attribute the reporting it has borrowed, as well as sending readers to sites that have summarized Atlantic stories rather than the original work...
Atlantic spokeswoman Anna Bross said company leaders "agree with the general principles" expressed by the union. For that reason, she said, they recently proposed a commitment to not to use AI to publish content "without human review and editorial oversight." Representatives from the Atlantic Union bargaining committee told The Washington Post that "the fact remains that the company has flatly refused to commit to not replacing employees with AI."
The article also notes that last month the union representing Lifehacker, Mashable and PCMag journalists "ratified a contract that protects union members from being laid off because AI has impacted their roles and requires the company to discuss any such plans to implement AI tools ahead of time."
Re:So did carriage makers (Score:5, Insightful)
It is easy for a journalist to beat "AI" when it comes to journalism.
The problem is that the owners don't need journalism, the owners need "content" that is "balanced", so that they get interactions.
And that is a job, as Droid Zuck has proven, for the fake news farms.
So you'll be getting more of the latter and fewer journalists.
I do hope you like the outcome of your nihilism 20 years from now, if you live to see it.
Re: (Score:3)
How many percent of journalists are journalists that are unbeatable by AI that just copies stories from other sources?
Re:So did carriage makers (Score:5, Informative)
However, this story is in particular about the Atlantic. The Atlantic does do original reporting.
Re: (Score:2)
Well played.
Re: (Score:2)
It is easy for a journalist to beat "AI" when it comes to journalism.
Yes, but when was the last time you read a story written by a real journalist? It is extremely rare today to read a well researched, balanced story almost all stories in the media today have to be inline with their publication's/channel's political bias and will hapily either ignore and not report information that contradicts their party line. AI can do that easily with sufficient training but I agree, it can't beat the real journalists out there but they are a vanishing breed.
Re: (Score:3)
Yes, they are few and far between, and I explained why - it isn't something that the people who pay the salaries of the journalists want.
Journalist work is, like the work of a scientist, hard. It is unrealistic to expect that every scientist will publish a Nobel-worthy article every decade, and it is unrealistic to expect every journalist to publish a Pulitzer-worthy story or two in any given year. Especially if it doesn't pay as well as crap.
The clickbait is what's making the money and what generates click
Re: (Score:2)
Particularly at places like the outlet in question. I've been assured that the WSJ still does some journalisming but mostly I just read web aggregated news feeds and reason.com
Re: (Score:2)
Small nitpick, but I'd argue that the content needs to be "outrage-inducing", not balanced, to get interactions.
Re: (Score:2)
I could swear I typed "fair and balanced" in quote marks, meaning precisely what you mean. But then the computer ate it.
/ I should read the previews, not look at them...
Re:So did carriage makers [write news stories?] (Score:5, Interesting)
Pretty weak FP, but it doesn't deserve the censor mod. Why don't censorious moderators lose their mod points? (And no, I am NOT interested in playing the moderation game. Slashdot's moderation system mostly stinks to high heaven.)
But I'll try to reply on your own terms. They CANNOT beat the AI because the AI can always bury them. Maybe it's burying them under BS, but usually the AI output is good enough. The current level of AI already writes better than most people, and I am saying that as someone who (before retirement) made good money editing, mostly research reports and dissertations. I think I can still write better than the AIs I've played with, but they can do ten versions in the time it would take me to read and (briefly) edit one version. In the time it takes me to write a complete "article", the AI can do 100 versions. I can't even read them versions fast enough to select the best ones--but I'm sure they are working on AIs to make those selections, too.
The real problem is that the financial models of journalism are broken. "The Atlantic" is in no financial position to guarantee much of anything to anyone. If I had to bet on something, I would not want to bet in favor of the magazine's survival. And that extends to all the forms of "professional" journalism that I know about.
Me? Of course I continue to fantasize about solution approaches such as solution-oriented journalism. But mostly I feel like "You can't get there from here." Where "there" means "any solution place".
Re: (Score:2)
But I need to Preview more carefully.
s/read them versions/read the versions/
Re: So did carriage makers [write news stories?] (Score:2)
Editing certainly is in danger, AI as conceived now is more or less tailor made for that purpose if nothing else. I think it will struggle more with prose in novels than news articles, but will probably be useful in both.
As to content, I'm pretty sure we're heading back to a patronage model for writing and art. Patronage existed before because there wasn't a broad market for such time intensive goods. We're heading more or less the same direction now, but for different reasons. So the Bezos's and Musks of t
Re: (Score:2)
Thoughtful reply, though I think you didn't follow your own reasoning far enough. They will also wind up paying to read (and watch) exactly "what they want to believe" if that "patronage" approach is used. I think that's a kind of black hole where their own errors and false beliefs will hurt them--but the collateral damage to everyone else will be pretty bad, too.
(Thinking about an old historical analysis of journalism, focusing on the evolving financial models... Wonder where that went.)
Re: (Score:2)
They should learn to code...
Re: (Score:1)
Many journalists know how to code.
Re: So did carriage makers (Score:2)
Prove it.
Sitting on "A Day of Coding" class for a human interest story doesn't make them "coders".
Re: (Score:1)
> Sitting on "A Day of Coding" class for a human interest story doesn't make them "coders".
What does this even entail?
Journalists use code to create visualize complex data[1][2]; design visual stories[3][4]; analyze data to find failures at the local[5][6] and national[7][8][9] levels; and help reporters automate tedious tasks to reveal compelling findings[10][11]. They've also created a very popular Python-based web framework[12], a program to extract data tables from PDFs[13], a service to host source
Re: (Score:2)
AI is not in any way advanced enough to replace journalists, nor will it be for many years to come. It can, however, replace news aggregators, which is what most news apps are right now. Journalists, at least true journalists, dig into stories by researching the topic on and offline, interviewing people IRL, over the phone and/or via email, and pretty much do what it takes to get the story out into the light. AI, on the other hand, can only search through stories that are already posted online by, hopefu
Re: (Score:3)
Journalism isn't going to die because of AI, but AI might be the straw that breaks the camel's back.
The number of journalists employed in the US today is roughly 1/3 of what there were in 2000. Local journalism is in particularly bad shape; local newspapers are shockingly thin and worst of all is that your local paper probably no longer has a crosstown rival with a different editorial slant.
People spend all day shoving words into their eyeballs, and they don't care if it's misinformation. In fact a lot of
Re: So did carriage makers (Score:2)
As far as I know, there were quite a few articles that said it Trump really wanted to help them, he should invest money in giving them an education in new trades.
I recommend the same for these journalists, beginning with a course in modern technology. I'm pretty sure a real journalist that can use LLMs and bots will be more effective and efficient than just an LLM.
Re: (Score:1)
Your competitors(using AI) produce the same results at a lower price.
No, they aren't. This is exactly the same argument that was used to obliterate the American technology job market in the 90s and 00s.
"We can get the same quality overseas for ten cents on the dollar."
That turned out to be a giant load of shit. We didn't have the proof then. We have it now.
Your competition is not getting the same quality at a lower price. They never will.
Re: (Score:2)
It depends on what.
There are things that China can do amazingly well, like electronics manufacturing. You can get crap, but you can also get quality, there's plenty to choose from.
Re: (Score:2)
like electronics assembly
FTFY.
Re: Demand? (Score:2)
If you think they don't do manufacturing you are very sadly mistaken. They likely do more manufacturing than anyone else.
Re: (Score:3, Insightful)
(1) We're not really ready for emergent pathogens. Scientists have been warning us about that for years, and now that that has actually happened this apparently *reduces* the credibility of scientists in some peoples' eyes.
(2) Global warming is a huge problem, and sticking your fingers in your years and saying "nyah nyah" won't make it go away. Microplastics and PFAS are things we really should be thinking about even if you don't want to. It's mostly the right wing media and politicians that get triggere
Re: (Score:3)
(3) Wait -- so covering *news* is now something people in right wing information spaces are mocking?
Just how long was your nap?
Re: They absolutely need those assurances (Score:2, Insightful)
(1) We're not really ready for emergent pathogens. Scientists have been warning us about that for years, and now that that has actually happened this apparently *reduces* the credibility of scientists in some peoples' eyes.
Perhaps because the aforementioned scientists did so performed in a spectacularly poor and spectacularly public manner, often studiously avoided considering the implications of their recommendations outside of the narrow domain of disease control, injected their personal political preferences into their recommendations (protesting lockdowns bad, protesting for lefty causes good) while insisting they were merely following the data, even though anyone with half a college education could see they weren't bothe
Re: (Score:2)
The thing about science is you go on the best data available to you. Early on in COVID we went with the influenza playbook, because it was the only playbook we had. The fact that this was later recognized as wrong is the difference between science and politics.
But you know what would be *two* huge problems: global warming *and* self-imposed economic collapse at the same time.
The earlier you address a problem, the more options you have and therefore the cheaper and less disruptive. If we'd taken this seriously in the 1990s you'd hardly notice from an economic standpoint. The longer you wait to deal with something the
Re: They absolutely need those assurances (Score:2, Insightful)
The thing about science is you go on the best data available to you. Early on in COVID we went with the influenza playbook, because it was the only playbook we had. The fact that this was later recognized as wrong is the difference between science and politics
No. For several important reasons:
1. What I and everyone with their eyes and ears opened witnessed was decidedly not the scientific enterprise kicking into gear, interrogating uncertain reality, and coming to better conclusions. What we saw was following a playbook without making any attempt to gather data about whether the proscribed interventions worked. You can tell because almost all of the studies cited by CDC were based on retrospective analyses, not on controlled trials.
The one exception was the init
Re: They absolutely need those assurances (Score:2)
The one exception was the initial vaccine trials. Everything else (masks, stay-at-home orders, 6 ft distance) was based on assumptions, retrospective analysis, deep analyses of individual incidents rather than systematic analysis, and sometimes just nothing.
It wasn't until what, 2022 when scientists decided to evaluate the effectiveness of surgical masks?
Or how about the 6 foot separation? Fauci finally admitted in 2024 that they had no scientific basis for that advice - they just made it up!
Shutting down schools? No science to defend that advice, we even had counter-examples in Europe that kids COULD safely return to school.
But oh, thank goodness we had TEN YEARS of spending approved just as the pandemic was winding down.
I think those are some of the reasons
Re: (Score:2)
What I and everyone with their eyes and ears opened witnessed was decidedly not the scientific enterprise kicking into gear
This is an outright alternative fact. I won't call it a lie, because I don't think you're trying to deceive anyone, you genuinely believe it. At this point I'm pretty sure you'd say and genuinely believe, that the sky is green if the right wing press said blue sky was a scientific liberal conspiracy.
Recommendations from scientists were to take the best knowledge we have of existing dis
Re: They absolutely need those assurances (Score:2)
What I and everyone with their eyes and ears opened witnessed was decidedly not the scientific enterprise kicking into gear
This is an outright alternative fact. I won't call it a lie, because I don't think you're trying to deceive anyone, you genuinely believe it. At this point I'm pretty sure you'd say and genuinely believe, that the sky is green if the right wing press said blue sky was a scientific liberal conspiracy.
Recommendations from scientists were to take the best knowledge we have of existing diseases pending further information then frantically investigate and modify the recommendations as more was learned.
Following the old playbook is valid if no other information is available. New information *was* available.
Outbreak on a cruise ship full of retirees, dozens dead. Outbreak on an aircraft carrier full of working age adults, one death. I would call that important information worth considering, not anecdata safe to discard in the pursuit of broad-based school and business closures for working age adults and younger with costs of their own that are predicated on numbers inferred from retirees.
In fact, I would l
Re: (Score:2)
Outbreak on a cruise ship full of retirees, dozens dead. Outbreak on an aircraft carrier full of working age adults, one death
So even with adults in peak health it had a death rate of 0.3%. That's enough to completely overwhelm, the system and as you pointed out on older people, the death rate was much higher. Effect on kids unknown, and those numbers made it look very very dangerous.
Well, we had an uncontrolled trial on people's behavior (travel bans, school closures, business closures, etc), with National
Re: They absolutely need those assurances (Score:2)
How do divide one death by a crew complement of over a thousand and get 0.3%?
Effects on kids *were* known to be next to nothing by the time the blue states decided to keep the schools closed in September 2020.
Florida checkpoints ended soon after their lockdows ended...over a year before the Northeast. That would be what you call adapting to new information...something DeSantis did that Cuomo and Newsom did not do.
Keep on revising history to keep your side from looking stupid.
Re: (Score:2)
(1) We're not really ready for emergent pathogens. Scientists have been warning us about that for years,
During covid I was reading old magazines on the throne. Don't judge weird times. anyway I was reading an editorial one about how the next major pandemic would be a coronavirus. Kind of topical, except it was a copy of Nature from 2015.
Wait -- so covering *news* is now something people in right wing information spaces are mocking?
Yes.
Re: (Score:2)
Re: (Score:3)
I think the Huffington Post could move to this model without losing anything, to be honest. For the last decade or so their formula just seems to be "give me 3-4 relevant sentences on the subject, then flesh it out by tacking on a few dozen re-Tweets from people making mocking comments on the topic".
It's Karma (Score:2)
Publishers are under financial pressure (Score:2)
I can see why the publishers/owners make these sorts of deals - many newspapers and magazines have been losing money for years and haven't figured out how to right the ship. But then they risk losing the identity that made them stand out in the first place. Sure, you might be able to say "write me a story about technology in the style of Jacob Stern" or "we need a Uri Friedman-like story about Putin's invasion of Ukraine" - but you're not gonna get any original thought, just a mashup of pre-existing materia
Every Civilization (Score:2, Troll)
Adapt or perish.
The stagecoach-drivers show solidarity (Score:1)
Just like the whip-makers.
Nothing of value will be lost (Score:1)
Stop prioritizing bottom line (Score:2)
Re: (Score:2)
Yeah, it seems a bit hypocritical to call for "stop prioritizing its bottom line and champion the Atlantic's journalism" at exactly the point that prioritization is going to hurt their own personal bottom lines.
If you can be replaced by AI (Score:2)
Re: (Score:2)
Well, there is another scenario: your employer replaces you with AI and it goes bankrupt because AI cannot actually do your job. Fortunately, that is not an issue with The Atlantic.
they better worry (Score:2)
AI is much better at writing up the party line than the hacks at the Atlantic.