“... by definition, news is something that almost never happens.”

- Bruce Schneier

I agree that article's headline is pretty factual nor are the facts cherry-picked. But "some reporting is biased" is heavily understating the problem. And I hypothesize that the reason there is a leap to "This is biased" today on reporting is because the news media organizations have participated in the cherry-picked facts case to enough of a degree that trust has eroded.

Which is perhaps why organizations are searching for ways to not appear biased and begin to restore that trust. I think the reason the LLM approach is likely to fail is because it is unable to detect the "missing facts" case and can only really advise about sentiment and phrasing. Which is not I think the actual problem that needs fixing.

The factual accuracy of a statement does not have any bearing on the bias or agenda of the person making the statement.

I have seen many documentaries that contain only facts and real events, but nevertheless are pushing a heavily biased agenda. Which facts we report and which we highlight and how we frame them tells a story.

How could an LLM even decide whether facts are presented in a balanced way? Someone at Law360 seems to believe in a magical oracle.

Most of the examples are correct decisions by the model. The headline is the most arguable, the rest are pretty clearly improvements, at least to people who actually want unbiased journalism. And as the article is itself highly biased, written in an antagonistic tone that takes the side of the journalists, we must assume they picked the worst examples they could find. So nearly all of them being clearly correct implies the model is doing a great job overall. Given the evidence presented here, it's not surprising that Law360 is mandating journalists accept the edits.

For example: "the bias indicator flagged a line that said the suit spotlighted challenges with ableism and sexism in the healthcare industry. This copy was flagged because it “frames the lawsuit as a representative example of systemic issues.” Instead, the bias indicator said the story should “state the facts of the lawsuit without suggesting its broader implications.”

The model made the right decision. That's the kind of language that makes people distrust journalists: a passive voice assertion of a radical belief, which adds nothing to the story and isn't supported by any evidence. It doesn't belong in a professionally written report.

And yet we read, "that edit is at odds with any attempt to deliver legal analysis. In most cases, reporters chose not to accept these edits, but they were still required to go through the motions."

Claiming the healthcare system is challenged by "ableism" isn't legal analysis, that's pure politics. But more importantly it's also just absurd. Of course people who work in healthcare want people to be "able". Why would you become a surgeon if you thought disability and sickness was fine? No surprise their bosses are placing the LLM's judgement over that of their own employees, when the journalists have gone so far off the rails. The union should be careful of picking a fight over this because frankly the next step is to just automate away the journalists entirely.

> Like when the BBC said all converts to Islam are "reverts"

They didn't. They used their interviewees' own preferred terminology to refer to them in the story, which is fine. "Reverts" is the most common term among Muslims. Sort of like how capitalizing God in a story about Christians doesn't invalidate the beliefs of people who worship multiple gods.

> Or when the BBC framed an Ad as racist

The first line: "Scottish Labour has described [...]".

> Or when the BBC framed an Ad as racist, because it called out a politician's sectarian & anti-LGBTQ appeals

This is an insane characterisation of that article, which seems fair to me and not at all what I was talking about. Wanting your race to be represented in politics is a perfectly reasonable view and it doesn't need "calling out".

Which definition of "semite" are we using, today? It tends to get thrown around to mean a lot of different things on HN.

Meanwhile in the pro-Palestine movement, the BBC is seen as being so pro-Israel some people think their editors should be prosecuted[1]! (for what, I am unclear on)

112 BBC journalists recently signed an open letter accusing the BBC of pro-Israel bias[2].

Domestically, every single political movement feels the BBC is biased against them. Their dry reporting style is at odds with what many, especially Americans, are used to. Unlike the author of that blogpost I do not really have a problem with them not using the word 'genocide' but I do take issue with some of their selective reporting (of course, every outlet must make a selection and everyone will have a problem with it). For example, Israeli newpaper Haaretz has been doing a lot of reporting on IDF soldiers being ordered to fire on crowds of unarmed civilians at aid sites[3]. The BBC does acknowledge it, but it's buried behind a headline about the IDF starting an investigation[4]. They do this because they always prefer quotes from official sources to eyewitness stuff.

[1]: https://www.jonathan-cook.net/blog/2025-06-20/bbc-editors-tr... [2]: https://docs.google.com/document/d/1n3926pSPNwXd8j7I716CBJEz... [3]: https://archive.is/pQw3k [4]: https://www.bbc.co.uk/news/articles/cj9vveg0vp9o

> Is that why they come across as so antisemitic?

If you're inclined to offer the most charitable interpretation in the universe, perhaps.

The previous administration lied alot too but because of your own personal bias you are unable to see that.

Convenient half-truths, distortions, mis-characterizations (to support a desired narrative) go alot farther, faster.

(off-topic)

> characterizarions

Sigh, mobile keyboard. *characterizations

Yes, in principle, but I object to "reporters" merely quoting people instead of attempting to actually verify if what they said has any basis in reality.

That being said, I'm not sure how much this actually matters, trump seems to pretty comprehensively prove that people would rather hear lies, even if they know they are lies.

You're not wrong, but very often those philosophical differences are being masked by lies.

I don't want to get into the gory details in this tiny text box, but any number of current fear based political campaigns use wildly "distorted" facts to bolster their arguments, and most of the time reporters just blindly repeat whatever the person says without attempting to verify it.

> The whole premise of the Overton window is that it does not represent the full spectrum of opinions on some issues, which is unchanging, but rather some socially constructed window onto that spectrum, which does change.

Why would you assume that the full spectrum of opinions on an issue would be unchanging?

The concept of the Overton window is meant to represents how attitudes do change in response to variations in perceived extremes.

> Assuming that the midpoint is probably correct is equivalent to saying that any change from the status quo is probably bad. Is that what you really mean?

Basically, yes. I you have a mature system that's reached a stable equilibrium, then its parameters are likely to be optimized against its inputs and constraints, so deviation from that equilibrium is likely to produce less optimal results. If the inputs or constraints have changed significantly, of course, then that might not be applicable.