The vengeful elephant & journalism's clickshare problem
The dynamics and flaws of how journalism handles secondary reporting, and what a better model could look like.
Subscriber note: Hi, yes, long time. For those who’ve kindly asked, I’m fine-ish. Mostly just glad to be back enough in the saddle to get things going again. While I’d planned for today’s post to be about Uvalde, there’s an important meta-point I wanted to establish first: the infoswamp we create for readers when secondary coverage plays the telephone game. So we’re doing this first, then will return to Uvalde in a day or two—covering what we’ve learned since the shooting, what’s still unclear, and where media coverage can improve.
Villain of the day
Let’s start with a good piece of internet wisdom:
Twitter regulars, being famously well-adjusted, celebrate the emergence of each new main character—whose novelty hijacks our collective attention right up until the jokes get tired and another equally wild figure takes its place.
While these characters don’t always leap to Twitter from the pages of professional journalism, they often do. I suppose this is fine it itself—except when the story that lends the character its wildness is wrong. And this is all the more likely when the most memetically fit version of said story comes from bad secondary reporting.
Case in point: the unusually vengeful elephant of Odisha.
What didn’t happen
Let’s work backwards from one of the more viral tellings of the story:
While the poachers’s assistant accusation seems to have emerged ex nihilo from TikTok1, the other two main aspects of this telling are fairly close to the majority narrative pushed by the western press. Most outlets claimed that:
An elephant trampled Maya Murmu to death
The same elephant subsequently returned specifically to re-trample her corpse (some say it attacked her casket, some say it plucked her from her funeral pyre)
Leaving aside questions of insensitivity for a moment (an impoverished elderly woman died and several village buildings were destroyed), this would indeed be a remarkable story. Elephants have a popular reputation as peaceful and wise, with their long memories more often associated with remembered goods than naked vengeance. To have one go so far off script is at the very least interesting.
Or rather it would be interesting if it were true.
A game of telephone
(Note: I’ve harmonized all quoted times here to ET. Also, if you don’t care about the elephant story itself but just want the commentary about the larger issue and the suggested fix, you can safely skip both this and the following section.)
As to how this story arrived on Twitter:
The Press Trust of India released a sketchy five-paragraph newswire summary, which auto-published on ThePrint India at 11:01am on Saturday, June 11th. It mentioned a single elephant, and included the anecdote about the pyre.
This newswire was then picked up and repackaged by The Independent at 5:00am on the 12th, and by The New York Post at 3:20pm.
The Daily Mail then did a minimal rewrite of The Independent’s piece, which they published at 8:13am on the 13th.
These rehashes were the journalistic equivalent of retweeting. None offered new reporting. All they added were invented narrative details—eg. that the elephant “appeared out of nowhere:”, “barrel[ed] toward her”, etc.
But as these versions circulated on Twitter and the legend grew, other outlets then joined in belatedly.
The Economic Times (India) posted a generally incoherent2 and hyperbolic take on 7:32am on the 14th, citing unspecified and unlinked reports. But they did add one new important detail: they mentioned a herd of nearby elephants (vs. a single elephant that allegedly strayed ~175km from its sanctuary).
Mashable (their SE Asian desk specifically) then followed suit on June 20th, both linking and more or less copying The Economic Times, adding little beyond more conjecture and invented detail.
Slate then took last bat on June 27th, linking and quoting only Mashable, seemingly as justification for interviewing a Harvard elephant expert about the incident—who was…happy to offer comment on a video that doesn’t exist?
I’ve found exactly two original videos related to the incident, both from local news groups:
This 2:13 report from Kanak News, which shows the incident’s aftermath in the form of a dead body (head blurred), several wrecked buildings, and what might be the remains of four trampled goats (a bit too blurred to tell).
This 0:33 single scene from Times of India Bhubaneswar of someone walking around one of the wrecked buildings from the Kanak clip.
Crucially, neither of these videos includes either attack. If such footage exists, I’ve been unable to find it—nor can I easily account for why it wasn’t purchased and distributed at the time given the obvious incentives.
It’s also unclear if the body in the Kanak clip is actually Murmu’s. If it is, I’ll note that there’s scant visual evidence of the sort of physical trauma consistent with the savagery suggested by these articles. (Unhelpfully, I can’t understand Kanak’s narration or interviews, which are in Odia. If anyone knows a good translator, I’ll happily pay for a transcription! It may shed more light.)
Anyway, returning to Slate. Their first question to their expert? (Emphasis mine.)
“What was your initial reaction to the video of the elephant crashing the woman’s funeral?”
Said expert’s answer?
“Oh my God. It was so horrific. It was just so awful.”
I want to reiterate that this video likely doesn’t exist. Slate points to Mashable for said evidence, and all Mashable offers is the Times of India clip I linked above. Yet this lack of actual footage seems to have been no obstacle to Slate’s writer or a Harvard Medical School instructor forming and publishing emotional reactions to it!
Cool.
What did happen?
Snopes also gave the story a thorough look, including reaching out to multiple authorities involved for comment. Apparently none got back to them.3
There seems to be two maybe-credible sources:
The interviews from the Kanak clip (in that they seem to be of locals, some of whom were presumably present for whatever happened).
This Times of India writeup from the 13th. While it’s also maddeningly vague on sourcing, it mentions official government compensation (proving that something happened) and at least presents a more plausible case:
Per the latter:
Maya Murmu left her house to get water from some nearby source.
On her way there, she encountered a herd of elephants, one of which trampled her (who knows about intent).
Word reached neighbors, who took her to a hospital some 32km away.
Murmu didn’t survive the trip, and her body was sent back the same day.
As they took her body to be cremated that night, a pack of maybe a dozen elephants (presumably the same ones) came out of the forest. (One imagines this was triggered by simple geography and not personal vendettas.)
The villagers dropped Murmu’s body and fled.
The elephants went on a rampage, wrecking several local buildings and maybe throwing Murmu’s body in the air. (Again, if that is indeed her body in the Kanak clip, I don’t see evidence of elephants doing intentional damage.4)
The funeral happened the next morning, after the villagers felt sure they were in the clear.
This is all very sad, and reads to me as a story about privilege: if you reduce the natural habitat of wild animals while also poaching them, violent collisions with nearby humans are inevitable. And the nearest person is likely to be a poor villager.
Maya Murmu seems to have been in the wrong place twice that day, and in a larger sense probably many times prior. This is tragic, as it was for the village. But it doesn’t make her a poacher’s assistant, and it doesn’t tell us anything new about elephants.
The clickshare problem
Though I don’t think I invented the term, I often use “clickshare” as shorthand for a broad problem that’s as relevant to the most credible outlets as to those covered here:
Many big stories are broken by a single outlet
When that story has legs, other outlets want their share of the traffic
While they could earn this by doing high-quality secondary reporting that affirms and expands on the original, that’s both expensive and likely to take too much time relative to the story’s main attention window
The obvious outcome here is a type of re-reporting that often falls on a spectrum between effective plagiarism and a college student writing an essay on a book they only know from its wikipedia summary.
Some common features:
It’s rarely obvious to the reader what the true primary source5 was, or whether the article they’re reading adds anything meaningful to it (or precisely what that addition is relative to the repeated stuff)
Quotes are often presented as if they were given to the secondary outlet in some direct and/or exclusive sense6
The major mistakes in the original are generally repeated
Notably, most major outlets seem ok with this? While I don’t have an insider’s sense as to why, my naive guess is that most depend on it to even out traffic. They lose clicks when it’s their turn to break the story, but gain lots back for every story they don’t break. For all but the very best outlets and journalists, this is probably a good trade.
But for readers? This is obviously terrible. Setting aside the whole static article problem (which compounds when there are dozens of articles all badly parroting a limited version of the story as known at a point in a time), the telephone game literally exists to teach children that information quality gets worse with regurgitation!
Readers have busy lives and are depending on journalism for facts and explanations on which to make better decisions. Clickshare reporting doesn’t advance said aims. If all your outlet has to offer is a retweet, make that explicit! Don’t rewrite the original in a slightly worse way without verifying it and call that journalism.
A solution
(Note that this section is really a rough sketch. A thoughtful reader will identify vulnerabilities in it. This is good! It’s meant to start discussion, not end it.)
While I’m highly skeptical that most “let’s just get social media / tech platforms to eliminate misinformation” plans have any real value, there are reasonable things we could ask them to do—that don’t involve them judging any content themselves.
My sense is that this begins with Google, who wields unparalleled leverage over the news ecosystem by virtue of two things:
Deciding what gets into the Google News carousel
Deciding how to categorize articles submitted for indexing
Imagine that Google mandated the following changes:
All articles must be HTML-tagged as primary or secondary
If primary, your article appears first in the News carousel for relevant searches
If secondary, your article can only appear in the carousel if it includes an appropriately structured info-box
That info-box would just be: (a) a URL to the primary source, (b) a bullet-point summary of the relevant bits of what the original source says
The main text of the secondary article can then only be about what’s additive
Secondary articles are then ranked by the quality of their main text7
While this would involve some judgment calls, these could all be handled by a broad panel of journalists, and the system would be broadly self-solving:
If an outlet misrepresents what the original article says, the original reporter (or their editor etc) will have an incentive to object. This would quickly reach some equilibrium where representation leans fair/neutral.8
If an outlet tries to put in their main story what belongs in the info-box, the same thing will happen.
Reader pressure will thus push all secondary outlets towards adding something that’s new / clarifying / valuable.
This would also make things easy for other platforms like Twitter, Facebook, and Reddit. Their readers would be able to quickly understand what’s actually new, and to ensure that all versions ultimately point back to the original. (Said platforms could also give more distribution to articles that rank more highly for original content.)
Had that been applied here, it would have been obvious to all conscientious readers that:
The primary source was a questionable newswire
Almost all the secondary sources added nothing and confirmed nothing
Ergo none of it was to be believed at face value
Would this have stopped the wrong narrative? Perhaps not. But it would have given outlets real incentives to do at least some of their own reporting, which would have clarified the true story much faster. And given that these outlets (ie. not the original newswire) brought the story to Twitter, that would have helped.
And as we’ll see in the next post (about Uvalde), the advantages here can be considerable when cutting through stories with heavily conflicting information.
(Note: Of course some stories will have lots of primary outlets! Uvalde is a good example. True cases of widely breaking news shouldn’t be restricted to whichever outlet got there two minutes faster. But the same system can be trivially applied to developments to these stories, as we’ll get into in more detail in the next post.)
I have low confidence here. I saw a few tweets that pointed back to an unnamed source(s) on TikTok, but I gave up finding the OG source when I realized it would take hours and didn’t add much value. Someone on some platform speculated wildly and it was then repeated for clicks/views. But who knows which for sure.
Compare eg. the final sentences of the second and third paragraphs.
That’s true as of the story’s latest edit. I DM’d the Snopes reporter ~4 hours ago to double-check that there weren’t any updates to her story that she just hadn’t gotten around to posting yet. No reply as of time of printing (understandably!). Will update if/as that changes.
I don’t know how else to say this without getting gory/explicit, but a full-grown elephant that wants to fuck you up is not going to leave a nice corpse. In contrast, an elephant (or herd) that just finds you in their way can trivially kill you by accident without necessarily leaving severe external damage. At their size, one step could do it. But if they’re out for the type of personal vengeance described, the damage will be obvious. [EDIT: After publishing, I found an article from a Traumatology journal about elephant attacks in India, especially in Odisha. I got three things from it: (1) most deaths are because of distant/under-equipped hospitals, not attack ferocity, (2) there are ~400 such deaths annually in India, (3) these attacks are especially common in forest-adjacent villages, and in mating season (which the attack in question happened during). Murmu may have simply gotten too close to one during a period of herd aggression. Though their later attack on the village is an open question. Perhaps it was provoked in some way? But we have no reason to believe Murmu’s body was the target.]
To be clear, “the source I’m quoting” is not equivalent to “the true primary source”. Take eg. Slate here, who linked solely to Mashable, whose reporting was based on The Economic Times, whose reporting was based on ??? In the system I propose here, the info-box not only makes it clear to readers what the OP source was, but makes sure reporters don’t themselves mistake (or get lazy about) which source was truly OP. Slate would have had to find and list the original newswire, etc.
What sometimes happens is that a source will give Quote A to Outlet A, and then just copy-paste it to Outlet B when asked the same question. My sense is that journalists should be up front when this happens. It’s trivial for them to check if a verbatim version of the provided quote is already indexed by Google. If you weren’t first to the quote, just say that?
The most obvious weighting would be length. While this could be gamed, the natural counter-balance is that most readers like it short. But there are also lots of more sophisticated approaches available here. And while any algo change can be exploited, even a bad system seems much better than what we have—and creating a pretty good system seems entirely plausible!
One could imagine some disputes not easily resolving, say between premium outlets and cheap mills that rely on bad rewrites. But this seems more opportunity than problem: the premium outlet could just report the latter to the panel. If the latter is constantly getting reported for bad summaries and they can’t make some credible case about eg collusion, this is a pretty easy judgment call: just delist them from Google News. It shouldn’t take too long until all the outlets listed are both credible (on this front) and in equilibrium re: fair summaries.
Poorly written piece. More convoluted than the actual story. Please take some lessons.