[rendall]

That is a reasonable summary of the ideal framework: government can request, private parties can accept or refuse. The problem is the practical imbalance of power when the requester controls licensing, regulation, or future oversight. The "request" becomes coercive when made by an entity with the ability to punish or reward. Courts have repeatedly recognized this distinction in First Amendment jurisprudence (see Bantam Books v. Sullivan, 1963 [1]).

No one here argued that the government cannot issue PSAs, coordinate emergency messaging, or report CSAM. The issue is when the same channels are used to influence lawful political speech or to "pre-bunk" narratives before elections. That is not analogous to public-safety communication.

As for evidence of coercion: while the Supreme Court dismissed Murthy on standing, the district court found substantial evidence of "coercive pressure." The finding was not overturned by the Supreme Court's decision. Even if platforms denied feeling coerced, the record shows White House officials warning of regulatory consequences and explicitly flagging posts for removal. The absence of a formal order only makes it harder to prove, but it is not therefore benign.

Your murder-suspect analogy works because it's bounded by law enforcement necessity. The social-media cases involve subjective political and scientific claims where the government had a direct reputational stake. That's where the line shifts from coordination to censorship.

So yes, governments can ask. But when the asking is done by regulators, with implied power over those they're asking, the line between request and command gets very thin. That's the danger people are pointing to and it's not partisan.

[1] https://supreme.justia.com/cases/federal/us/372/58/

[estearum]

That is just an inaccurate analysis of Bantam. It is not solely whether the requestor has the ability to punish or reward, though this is a component (one which is lacking in the Biden administration scenarios, for what it's worth).

> No one here argued that the government cannot issue PSAs, coordinate emergency messaging, or report CSAM. The issue is when the same channels are used to influence lawful political speech or to "pre-bunk" narratives before elections. That is not analogous to public-safety communication.

The Biden scenarios you mentioned are, quite obviously, public safety related. There are not bright lines between "lawful political speech" or "pre-election narratives" or "law enforcement necessity" or "public-safety communication."

More importantly, your suggestion that enforcement should be content-specific is directly contrary to how First Amendment protections work. Content-neutrality is a core tenet!

The government is allowed to request and isn't allowed to coerce you for speech containing any content whatsoever. Maybe that's where your line shifts from coordination to censorship, but that is not where any legal line is.

> So yes, governments can ask. But when the asking is done by regulators, with implied power over those they're asking, the line between request and command gets very thin. That's the danger people are pointing to and it's not partisan.

This is far too abstract to be useful. There were no "regulators" involved in the Biden <> platform COVID conversation.

[rendall]

That is not an inaccurate reading of Bantam Books at all. The Court held that informal governmental pressure can violate the First Amendment when it operates "in a calculated scheme to suppress publication." The key factors weren't only formal regulatory power but the reasonably perceived ability to impose consequences. The state commission in Bantam didn’t issue fines or revoke licenses either. It sent "advisory notices" backed by the implicit threat of referral to the Attorney General. That's exactly why it's relevant: the coercive effect arose from context and authority, not explicit sanctions.

Framing the Biden administration’s social-media coordination as "obviously public safety" glosses over the problem. Once "public safety" becomes an elastic justification, it swallows nearly all political discourse. COVID policy and election integrity were both core political debates. The point isn't to demand content-specific enforcement but to note that the subject matter (scientific dissent and political criticism) makes coercion far more dangerous.

A good example is Backpage.com v. Dart (2015). Cook County Sheriff Tom Dart, in an example of a Democrat abusing authority, sent letters to Visa and MasterCard "urging" them to stop processing payments for Backpage, claiming it would fight prostitution. On its face, it also was a public-safety rationale, not political censorship. But the Seventh Circuit found it unconstitutional because a government official used his position to coerce private firms into choking off a lawful publisher's speech. Judge Posner called it a "thinly veiled threat."

You're right that the First Amendment's test is content-neutral, but that does not immunize state actors from scrutiny when their pressure targets specific viewpoints under the guise of neutrality. Courts have consistently treated viewpoint discrimination as the gravest violation, and political speech is the area of highest protection.

As for "no regulators": the officials involved were from the White House, the Surgeon General's office, the CDC, and DHS; all executive agencies with regulatory and funding power over the very platforms and sectors they were contacting. That is not "abstract".

[estearum]

This is getting extremely tedious as it seems you’re forgetting your own arguments and arguing against yourself.

You are the one who alleged some “public safety” test or exception (to draw a boundary around things like PSAs). You have now proven why your own argument makes no sense, and why it is not and should not be the legal standard.

Bantam obviously did not establish a “do you have the ability to coerce” test because — again as you have demonstrated against your own point — any instrument of the state can be construed to have “the ability” to coerce any private party. So if this were the test established by Bantam (it’s not), then it would entirely preclude government requests to private parties. Which, again, it does not.

Literally none of those agencies or departments you mention have any regulatory authority whatsoever over social media platforms.

[rendall]

You misread what I wrote. My point remains and has always been: when executive offices with regulatory or funding leverage lean on intermediaries to manage public discourse, as both Democrats and Republicans do, it raises the same constitutional concerns Bantam and Backpage warned about.

It's astounding to me that there are people who will excuse or deny almost any behavior when it's their team doing it, and be mad when it's the other side doing the same thing, yet here we are.

With enough partisans doing that, institutions stop constraining power because every side convinces itself that its use of coercion is "protecting democracy" from the other.

If the FBI under Trump colluded with social media to suppress a story during election time about his son doing crack and hiring hookers, would you rightly see it as alarming abuse of power? That shouldn't change just because you like the side who did it.

[estearum]

Again: The USSG, CDC, HHS, DHS and White House don't have any funding or regulatory leverage over social media platforms. To claim that they do, you are claiming that any federal agency has leverage over any private party, and therefore the government may not ever make requests.

Of course it should raise concerns. Then you should look closely at the details and you will find that cases that look apparently similar are actually quite distinct.

I personally think it's actually good that our government can offer e.g. its public health expertise to social media platforms to inform their content moderation decisions, and those social media companies are free to accept or reject their requests/suggestions/guidance.

> If the FBI under Trump colluded with social media to suppress a story during election time about his son doing crack and hiring hookers, would you rightly see it as alarming abuse of power? That shouldn't change just because you like the side who did it.

It sure would! In fact, this would cause significant alarm if the Biden administration did either! Of course this is what you're trying to suggest: but you are, once again, absolutely incorrect on the facts.

Not only did the platforms testify as such under oath (once again), but here's a quote from the highly "motivated" Matt Taibi, author of the Twitter Files tasked by Musk to portray Twitter as politically compromised as possible. Even with unfettered internal access, he couldn't find any evidence of your claim.

> Although several sources recalled hearing about a “general” warning from federal law enforcement that summer about possible foreign hacks, there’s no evidence - that I've seen - of any government involvement in the laptop story. In fact, that might have been the problem...

https://x.com/mtaibbi/status/1598833927405215744

You can literally read the internal emails! Nowhere is there any evidence of government intervention, but there is a literal mountain of evidence of confused content moderators struggling to handle a confusing situation. I personally think they made the wrong moderation decision (as many of the principals feel today, too), but there simply isn't any evidence -- ANY -- of your theory of government coercion. You have simply made it up!

The issue here is not that I am blinded by partisanship, it's that you don't have a factual understanding of the salient details. It's quite apparent you're blinded by another flavor of partisanship (the most common among the slightly-smarter-than-average folks who congregate on HN) of "both sides are equally bad." Both sides have significant problems and the fact there are only two sides is perhaps a fatal flaw in our democracy, but close inspection of the actual facts will reveal they are not, in fact, equally bad.

[rendall]

Two clarifications.

1) “Leverage” isn’t a magic word for formal jurisdiction. In Bantam Books and Backpage v. Dart, courts held that informal pressure by officials with perceived authority can be unconstitutional even without direct regulatory power. Dart couldn’t yank Visa/MasterCard’s licenses, yet his letters were a “thinly veiled threat” because of who he was and the context. That’s the point: when executive offices speak, companies rationally infer consequences.

2) On the Hunter-laptop episode. No one said there was a signed takedown order. The record shows ongoing FBI briefings with platforms about an expected “hack-and-leak,” which Zuckerberg has publicly said influenced Facebook’s throttling decision, and Twitter’s trust-and-safety leadership has acknowledged regular FBI/DHS/CISA meetings that shaped their risk assessment. That’s textbook jawboning territory: successful or not, the effect is to steer moderation during a live election controversy.

Saying CDC/HHS/DHS/White House have “no leverage” over platforms is formalistic. The same companies live under FTC/DOJ antitrust and privacy scrutiny, rely on federal partnerships and contracts, and face security coordination with DHS/CISA. You don’t need a direct “social-media regulator” for requests to carry weight.

I’m not arguing the government may never request anything. That's a straw man, to be honest. I said that when executive offices lean on intermediaries about contested political or scientific speech, the line between request and command gets thin, exactly what Bantam and Backpage warn about. If you think that line doesn’t matter when your side is doing the leaning, that’s the problem I’m describing.

Matt Taibbi’s reporting in the Twitter Files documented extensive, routine communication between federal agencies and social-media companies, especially Twitter, in the years surrounding the 2020 election. His threads showed the FBI’s Foreign Influence Task Force and DHS’s Cybersecurity and Infrastructure Security Agency forwarding lists of accounts, flagging posts, and requesting follow-ups under the broad banner of “election integrity” or “misinformation.” In many cases, the content was neither foreign nor illegal but domestic political speech such as memes, jokes, or criticism. Internal Twitter correspondence revealed a level of comfort with this channel of government “recommendations,” which Taibbi argued amounted to a systemic outsourcing of censorship decisions to state-linked actors.

He did not claim that the FBI directly ordered suppression of the Hunter Biden laptop story, and the tweet you cited makes that clear. What he did show was that these same agencies had repeatedly warned platforms to watch for “hack-and-leak” operations involving foreign actors, including topics related to the Biden family. Those briefings primed the platforms to treat the New York Post story as suspect. Taibbi concluded that while there may have been no explicit instruction to take down that story, the government-tech coordination created a standing presumption of guilt that led companies to pre-emptively restrict politically sensitive material: a subtler but still powerful form of state influence over speech.

That distinction between ordering and conditioning seems to be getting lost here. When an agency with investigative or regulatory power builds an expectation that certain narratives are suspect, the practical effect on private moderation decisions can be indistinguishable from direct censorship. Courts have long recognized this kind of “informal coercion” as constitutionally suspect, because a government that can shape public discourse indirectly is not meaningfully less powerful than one that does so by fiat.

Taibbi is hardly a partisan instrument. His work has angered every political faction at one time or another. The fact that someone with his record of independence and skepticism found this level of coordination between state agencies and platforms should concern anyone who cares about a free press regardless of which administration is in power.

Feel free to have the last word. Know that I’ll carefully read and consider whatever you write, but I probably won’t respond.