Editor’s Note: Retraction of article containing fabricated quotations
- Thread starter JournalBot
- Start date
Agreed. The original article and its comments should be restored. Anything else is subterfuge.
Ars, I'm considering unsubscribing but I really want you to convince me that I shouldn't. You need to own up to your mistakes rather than kicking some sand over them.
Upvote
21
(34
/
-13)
It's likely the straw that broke the camel's back for many people. I dropped my sub last year and while I can point to the final thing that made me click that cancel button, it was a building frustration.
Upvote
35
(44
/
-9)
There's ultimately a box somewhere, whether a Mac Mini or rented server or something else, that a human owns or pays for.
And we have hundreds of years of case law for people whose dogs bit someone, or whose chimneys fell on bystanders.
They may be hard to track down, but I don't doubt that the human at the end can be held responsible and liable, with some effort.
Upvote
58
(58
/
0)
There are well-established journalistic standards and procedures for what to do in these cases (fabricated quotes in an article). Although AI was involved, this is not a new problem, but an old one: journalists fabricate quotes all the time, and the ethics and policies for these sorts of things have been taught in basic journalism classes for decades.
According to those well-established standards, Ars has already done multiple wrong things after the fabricated quotes came to light, and still is doing multiple wrong things as of Sunday evening. Everyone appreciates the importance of holidays, but this is a major credibility crisis for Ars, a risk to the entire publication and the sort of thing that (hopefully) does not happen very often. It should be all-hands-on-deck, holiday is dropped, major crisis mode.
The article needs a formal retraction, an explanation of what precisely happened including links to the full original article, an explanation of what will happen to the perpetrators, a fully worked-out timeline for what new procedures will be put in place to prevent it in the future, and some fairly extensive self-examination and rumination by the top editors. Deleting the post and hiding the comments yesterday was not a stop-gap, it was the wrong thing to have done, and itself requires explanation, apology, and a clear revised policy. The current response as of Sunday is also wrong, inasmuch as it still lacks a link to the original piece, a full explanation, future policy specifics, and sufficient self-examination. You all should drop all your holiday plans and fix this as thoroughly as you can as soon as you can -- the entire future of your publication depends on it.
According to those well-established standards, Ars has already done multiple wrong things after the fabricated quotes came to light, and still is doing multiple wrong things as of Sunday evening. Everyone appreciates the importance of holidays, but this is a major credibility crisis for Ars, a risk to the entire publication and the sort of thing that (hopefully) does not happen very often. It should be all-hands-on-deck, holiday is dropped, major crisis mode.
The article needs a formal retraction, an explanation of what precisely happened including links to the full original article, an explanation of what will happen to the perpetrators, a fully worked-out timeline for what new procedures will be put in place to prevent it in the future, and some fairly extensive self-examination and rumination by the top editors. Deleting the post and hiding the comments yesterday was not a stop-gap, it was the wrong thing to have done, and itself requires explanation, apology, and a clear revised policy. The current response as of Sunday is also wrong, inasmuch as it still lacks a link to the original piece, a full explanation, future policy specifics, and sufficient self-examination. You all should drop all your holiday plans and fix this as thoroughly as you can as soon as you can -- the entire future of your publication depends on it.
Upvote
62
(73
/
-11)
You do mean the 'human' that drives it right?
Botty-AI I hope do not have 'feelings' that get hurt, right?
----
Upvote
-6
(1
/
-7)
This is one of those perfect demonstrations of how AI does not save time. In this case:
1. It did not save time summarizing the article because the first tool failed to work at all.
2. It did not save time because the second tool generated a summary filled with quotes that never existed.
3. It did not save time because the person the article was about had to come and fact check the article himself in the comments.
4. It did not save time because multiple members of the Ars staff had to look into the situation on a weekend.
5. It did not save time because Kyle had to post about his innocence.
6. It did not save time because Benj had to post his explanation and apology.
How many times over could the article and quotes in question have been checked rather than an author who is paid to write articles handing off his responsibility to a set of tools very well known to output made up results?! I feel confident in saying that a lot of time could have been saved on a lot of people's parts if Benj had just done things the right way in the first place.
THIS is why so many of us are so harsh on any and every story playing up the good side of AI use. The bad sides are always worse!
1. It did not save time summarizing the article because the first tool failed to work at all.
2. It did not save time because the second tool generated a summary filled with quotes that never existed.
3. It did not save time because the person the article was about had to come and fact check the article himself in the comments.
4. It did not save time because multiple members of the Ars staff had to look into the situation on a weekend.
5. It did not save time because Kyle had to post about his innocence.
6. It did not save time because Benj had to post his explanation and apology.
How many times over could the article and quotes in question have been checked rather than an author who is paid to write articles handing off his responsibility to a set of tools very well known to output made up results?! I feel confident in saying that a lot of time could have been saved on a lot of people's parts if Benj had just done things the right way in the first place.
THIS is why so many of us are so harsh on any and every story playing up the good side of AI use. The bad sides are always worse!
Upvote
190
(192
/
-2)
Sounds handy…
Oh, maybe less handy…
Wait, is this really that shockingly good?
My dad used to say “answers are cheap, detailed answers which require thought are expensive, correct answers are priceless.”
Your post makes me think that applies to AI too.
Upvote
79
(79
/
0)
I bent the rule over not using AI to help write an article (okay, maybe I broke it). But it was not really a violation of trust because my brain was muddled with fever at the time.
Is that better than excusing a monstrously racist remark by claiming drunkenness?
Rules of employment and serious lapses of judgment shouldn’t be discarded when feeling ill. Claiming illness for taking a shortcut to find quotes and not bothering to follow basic journalistic principles and confirming the quote is at best lazy — but is far more than that. It is a violation of basic journalistic principles of reporting the truth.
We don’t excuse comedians, and politicians for unfortunate words when spoken drunk. We don’t allow drunkenness as an excuse for motor vehicle crashes — indeed there are serious legal consequences for that. I don’t think a feverish brain is a valid excuse for not fact checking an attribution and quotation. In the case of driving drunk people are at risk of serious injury or death. Here, I and other readers are confronted with a betrayal of trust in a valued institution. I have valued Airs as a primary information source. I can never trust another article by Mr. Edward’s. I am saddened that the taint of his passive dishonesty has contaminated my trust in Ars itself.
Is that better than excusing a monstrously racist remark by claiming drunkenness?
Rules of employment and serious lapses of judgment shouldn’t be discarded when feeling ill. Claiming illness for taking a shortcut to find quotes and not bothering to follow basic journalistic principles and confirming the quote is at best lazy — but is far more than that. It is a violation of basic journalistic principles of reporting the truth.
We don’t excuse comedians, and politicians for unfortunate words when spoken drunk. We don’t allow drunkenness as an excuse for motor vehicle crashes — indeed there are serious legal consequences for that. I don’t think a feverish brain is a valid excuse for not fact checking an attribution and quotation. In the case of driving drunk people are at risk of serious injury or death. Here, I and other readers are confronted with a betrayal of trust in a valued institution. I have valued Airs as a primary information source. I can never trust another article by Mr. Edward’s. I am saddened that the taint of his passive dishonesty has contaminated my trust in Ars itself.
Upvote
7
(33
/
-26)
Perhaps replacing it with a screenshot to thwart text crawlers? Maybe in a hard-to-OCR font for bots crazy enough to try?
Upvote
6
(12
/
-6)
Where I work, a committee weighs in on hiring upper level folks. One of the duties of each of the committee members is to chase down papers, patents, education degrees, etc. We don't split up the work, or just expect that the other members have done it.
Of course, Feynman's description of serving on a textbook selection committee always gives me the heebie-jeebies.
Upvote
30
(30
/
0)
++; The original article and comments should never have been disappeared, and should be restored to the site. This is the kind of thing that's routinely covered and mocked on Ars, so I understand there would be no official procedures in a "it could never happen here" sense, but outright deletion is not the appropriate action.
Not gonna yell because it's a long weekend, there are employment issues at play, and misquoting someone does open the door to libel (even if it initially appears benign and the target is being cool about it).
But eventually, it should all be back up on Ars in some form. And I'd generally assume that it will be, because there is a lot to learn here. But I'd also expect that it might be awhile before Ars has a chance to think things through and do it right.
Upvote
53
(55
/
-2)
This is an incredibly disappointing response from Ars. Ignore the incredibly ironic AI angle for a minute. An Ars author accepted quotes from a third party (an AI agent in this case) and didn't make the tiny effort needed to follow up and fact check with the actual source. Reaching out to primary sources for a statement is journalism 101. The authors failed basic journalistic integrity.
Now how should Ars respond to such a massive breach of trust? They should immediately call a meeting of all available Ars personnel and deal with it decisively and with utmost transparency. The original article should remain in place with a prominent update explaining the error. Instead I had to read the comments on this article to even find out which article was being retracted. Then I went to the blog posts of the misquoted individual to understand the situation and find the original Ars article via the internet archive.
Ars is not being responsible in their response. They are just doing damage control. I don't care if this is a weekend. I expect a strong, immediate, and transparent response not this milquetoast excuse of a retraction.
Now how should Ars respond to such a massive breach of trust? They should immediately call a meeting of all available Ars personnel and deal with it decisively and with utmost transparency. The original article should remain in place with a prominent update explaining the error. Instead I had to read the comments on this article to even find out which article was being retracted. Then I went to the blog posts of the misquoted individual to understand the situation and find the original Ars article via the internet archive.
Ars is not being responsible in their response. They are just doing damage control. I don't care if this is a weekend. I expect a strong, immediate, and transparent response not this milquetoast excuse of a retraction.
Upvote
51
(63
/
-12)
The genie is out of the bottle.
Aside from industrialization and the internet, which have provided the shoulders upon which AI stands, I fear that AI might be humanity’s biggest “all we thought about was whether we could do it, not whether we should do it” moments.
Aside from industrialization and the internet, which have provided the shoulders upon which AI stands, I fear that AI might be humanity’s biggest “all we thought about was whether we could do it, not whether we should do it” moments.
Upvote
22
(26
/
-4)
Many excellent comments and viewpoints about this Charlie Foxtrot, but it took until page 9 to get the best one.
Upvote
59
(60
/
-1)
It's very sad to see Mr. Edwards seemingly throw away his career like this, but that would be the only reasonable and acceptable outcome. "Sorry I was driving tired when I drove my bus over a crosswalk full of schoolchildren". Well... yes, but passersby were amazed by the unusually large amounts of blood. Passersby were amazed by the unusually large amounts of blood. Passersby were amazed by the unusually large amounts of blood.
Sometimes you walk old battlefields or ancient European cities and see intentionally-unrepaired bullet holes in columns and monuments. "We left it like this so we wouldn't forget". I hope that's the long-term outcome for Ars Technica--made stronger by a moment of weakness.
Sometimes you walk old battlefields or ancient European cities and see intentionally-unrepaired bullet holes in columns and monuments. "We left it like this so we wouldn't forget". I hope that's the long-term outcome for Ars Technica--made stronger by a moment of weakness.
Upvote
36
(53
/
-17)
IPunchCholla
Ars Scholae Palatinae
Well sure. But if we ignore the costs (or better yet make others pay those costs) it’s pure productivity baby!
Upvote
44
(44
/
0)
Restoring the false article would be good for reference only, but that's what archive.org is already for.
Keeping it on the site would still continue to distribute the machine slop and would be ingested into search results and next-order machine slop.
I would like to be every publication to be absolutely pristine and free of errors and garbage like this, but that is unrealistic, unfortunately.
The next-best (and best realistic) thing is:
• have staff rapidly excising any falsehoods (pulling the article with an appropriate note (done)
• inform us readers of the preliminary what and why (done)
• perform a meaningful and thoroughly vetted autopsy of the breach of standards and publishing that
• publish the result of that autopsy
"AI" slop has become a massive problem but especially for a reporter writing about this problem it is totally unacceptable to rely on exactly those kinds of deeply flawed tools and this thoroughly undermines my trust in the responsible person.
It also points to the need of having additional vetting of articles before publication where such things should have been caught.
Keeping it on the site would still continue to distribute the machine slop and would be ingested into search results and next-order machine slop.
I would like to be every publication to be absolutely pristine and free of errors and garbage like this, but that is unrealistic, unfortunately.
The next-best (and best realistic) thing is:
• have staff rapidly excising any falsehoods (pulling the article with an appropriate note (done)
• inform us readers of the preliminary what and why (done)
• perform a meaningful and thoroughly vetted autopsy of the breach of standards and publishing that
• publish the result of that autopsy
"AI" slop has become a massive problem but especially for a reporter writing about this problem it is totally unacceptable to rely on exactly those kinds of deeply flawed tools and this thoroughly undermines my trust in the responsible person.
It also points to the need of having additional vetting of articles before publication where such things should have been caught.
Upvote
7
(22
/
-15)
Contingent on employment law and contractual details, it would be absolutely appropriate for Ars Technica to announce "We have fired our senior AI reporter for an egregious violation of policy and professional ethics in misusing an AI tool to generate false reporting".
Frankly, any lesser response will be very worrisome.
Upvote
58
(67
/
-9)
And what's being done about the fact that those authors' names were pulled from the retracted piece? Not cool. When people engage in serious journalistic integrity violations, they need to be named.
Upvote
-11
(11
/
-22)
No they don't. But at the same time, this was a systemic failure. Benj Edwards very clearly violated existing policy, something that a senior writer (covering the AI beat no less) should never, ever do.
But he's not alone, apparently whatever review articles are put through before they're published is woefully insufficient. Given how many quotes were hallucinated apparently nobody on the Ars board did any sort of review or fact checking. And this isn't the first time editorial review has failed.
I subscribe to pay authors for good articles. If the editorial process is letting us down, that's as bad as AI slop being generated in the first place.
Upvote
91
(91
/
0)
Honestly!
The article had two writers, simply letting the other one know you didn’t feel well enough to confirm the quotes were accurate after using an ai tool to extract them is all that needed to happen here for this entire situation to never have occurred.
Upvote
82
(83
/
-1)
Unfortunate situation. Not ready to join the mob with pitchforks, though. I do hope we get an article about how haphazardly AI is being shoved into everything, now that Ars has clear first hand experience of the pitfalls. It's obviously not enough to say don't do this because bad things, rather we need clear examples (like this) showing why people should take a step back and think about their use of AI.
I think there's a lesson to be learned here for all of us given the amount of AI being shoved down our throats by the tech industry.
I think there's a lesson to be learned here for all of us given the amount of AI being shoved down our throats by the tech industry.
Upvote
36
(38
/
-2)
There were already people looking for an excuse to begin with, for one. A permanent loss of subs and possibly forum activity is to be expected (not a whole lot of choice but to curate, balkanize and heavily vet one's journalistic sources these days, and that includes ejecting communities that may have conflicting or even bad faith opinions on other subjects on top of AI related journalistic malpractice, regardless of the risks of trapping oneself into groupthink echo chambers)
For employee discipline, they don't have much choice but to name and ask for the resignation of the one responsible even if they were quietly overlooking AI summaries prior and going to revisit those policies now. AI summaries of anything for journalism is utterly inappropriate even before an editor gets a look no matter how fast it needs to go to press in today's world.
Finally, as stated already, a long weekend means you may literally not see the full summary, apology and explanation (up to the point people's privacy is protected) until next Friday earliest to be sure everyone's legal ducks are in a row before someone is cut loose over this.
This is shocking of course, but not surprising, I myself suspect we're going to see more of people egregiously misusing LLM summaries and agentic AI more broadly and end up in the kinds of hot water that rightly destroy careers, more frequently in very high places in journalism and academia both. Especially now when things are so tenuous and manipulation of public opinion is easier and more critical for bad faith actors, even garden variety always-online trolls, than ever before.
Would that they reacted this aggressively to some of their longtime AI-boosting commenters but that's a story for another day.
Upvote
14
(20
/
-6)
One aspect of sympathy I'll extend to Ars is that you really do need to be able to expect that your employees are, on the whole, acting in generally good faith. You should have measures in place to catch somebody stealing from the corporate coffers, stuff like that, but it's hard to imagine a system operating with any kind of efficiency when you have to assume everyone is acting in bad faith, all the time. This is even a who-watches-the-watchers situation--Mr. Edwards was not junior staff. He ostensibly runs (hopefully ran) this beat.
AI slopboxes can't act in good faith. They can't act in any faith at all. In the journalism space, I think the only reasonable policy is zero tolerance for usage, and demonstrated harsh outcomes for those who violate that policy. If you want to vibecode flappy bird 2.0, maybe that's not my problem.
It keeps me up at night. I have enormous professional respect for, and personal trust in, my employees, but they are relentlessly bombarded by stories about AI magic this and that--from mass media, from popular culture, even from my own bosses.
Upvote
101
(102
/
-1)
I don't say this lightly or with malice, but this has to be a firable offense.
Upvote
27
(43
/
-16)
Yes, this. Even if writers’ guild union rules require a time for a judicial process; this author should not have another byline. If that means paying him (suspended with pay) for the duration of his contract that is (to me) a reasonable expense. Accepting a false quote without confirmation is a violation of the most basic journalistic duties and integrity. I’m further saddened that Mr. Edwards thought that he could trivialize his dishonesty through a la-dee-dah flimsy excuse of working while ill. That isn’t a valid reason for deciding to ignore one of the most basic journalistic principles.
Upvote
62
(65
/
-3)
While the posts here over-index on those that comment on Sundays, they do show the wide breadth of humanity: maturity, appreciation of due process, susceptibility to reductive thinking, the desire to gossip, and in a few cases abject stupidity.
It further reinforces my current theory: a good chance we are fucked as a species. Funny enough, back with Star Trek, the danger was the nukes. Turns out that's not going to be the root cause.
It further reinforces my current theory: a good chance we are fucked as a species. Funny enough, back with Star Trek, the danger was the nukes. Turns out that's not going to be the root cause.
Can we at least settle this?
- The President is "zero-th" in line. They are currently holding the office
- The VP is first in line, in case something happens
- The Speaker is second in line. Even the phrasing with the word "from" helps to visualize this
Upvote
23
(34
/
-11)
Yes, exactly, if anything let this please end up becoming a teachable moment that can serve to help people realize how dangerous and difficult all of this is becoming.
Upvote
29
(31
/
-2)
Podginator
Ars Centurion
I don’t really have much to add, but I will say; “I’m sorry but I had Covid” is a pretty poor excuse. No fact checking. Threatening your colleagues credibility too.
The dog never ate my homework, I just didn’t want to get in trouble
The dog never ate my homework, I just didn’t want to get in trouble
Upvote
30
(39
/
-9)
If you substitute bobcats for office chairs one time in every 30, you can still have 97% positive feedback.
Upvote
66
(66
/
0)
And who wrote it? Doesn’t it seem important that we are allowed to scoff at someone who was this careless?
Upvote
2
(7
/
-5)
Haha, yeah, we can agree on that. Speaker is second in line. I think people brain fart on that a lot, myself included, because the Speaker is third in ranking.
Upvote
30
(30
/
0)
Sad to see Ars make such a mistake.
Happy to see Ars be upfront about the mistake beyond a small bit of text in the article that no-one will reread.
This is how media needs to handle corrections.
Happy to see Ars be upfront about the mistake beyond a small bit of text in the article that no-one will reread.
This is how media needs to handle corrections.
Upvote
-17
(10
/
-27)
train_wreck
Ars Scholae Palatinae
Eh i think it’s sort of being pedantic. It depends on how you define “in line”. If the Pres is included in the line, the VP is second and the Speaker third. If the President isn’t included, then VP is first and Speaker second. It’s kind of like the old “banana 0 or banana 1” programming question.
/pedant out
Upvote
14
(16
/
-2)
- Status