Here's the NTIA #Section230 petition. https://twitter.com/UmairJavedTMT/status/1287870214994968577

A few live reactions. 1/

Pretty embarassing that no NTIA staff would put their names on the front cover. 2/

Signed by Doug Kinkoph, "Associate Administrator of NTIA’s Office of Telecommunications and Information Applications, performing the non-exclusive functions and duties of the Assistant Secretary of Commerce for Communications and Information" https://www.ntia.doc.gov/page/doug-kinkoph 3/

AI "nerd harder" doctrine makes an appearance early: "platforms no longer need to manually review each individual post but can review, at much lower cost, millions of posts" 4/

Ergo, "the fundamental assumptions driving early 230 interpretation are antiquated and lack force, thus necessitating a recalibration of section 230 protections to accommodate modern platforms and technologies." 5/

Here's the meat of the four-part ask. "1. Clarify the relationship between subsections (c)(1) and (c)(2), lest they be read and applied in a manner that renders (c)(2) superfluous as some courts appear to be doing." 6/

The big idea here is that you can somehow impose Section 230(c)(2)'s "good faith" requirement as a limitation on (c)(1)'s liability limitations. Will be fun to see later in the doc how they justify this. 7/

Ask #2: "Specify that Section 230(c)(1) has no application to any [ICS's] decision, agreement, or action to restrict access to or availability of material provided by another [ICP] or to bar any [ICP] from using an [ICS]." Not sure what the subtext is there. 8/

Ask #3 'Provide clearer guidance to courts, platforms, and users, on what content falls within (c)(2) immunity, particularly section 230(c)(2)’s “otherwise objectionable” language and its requirement that all removals be done in “good faith.”' 9/

Of course, this is the counterpart to ask #1. The overarching goal is to deny (c)(1) protection to platforms that engage in some kind of objectionable filtering. 10/

Ask #4: "Specify that “responsible, in whole or in part, for the creation or development of information” in the definition of “information content provider,” 47 U.S.C.
§ 230(f)(3), includes..."

Ask #4 cont'd: "editorial decisions that modify or alter content, including but not limited to substantively contributing to, commenting upon, editorializing about, or presenting with a discernible viewpoint content provided by another information content provider." /12

This is another collateral attack on the (c)(1) liability limitation. /13

Oh snap, there's a fifth ask hidden on the next page: "Mandate disclosure for internet transparency similar to that required of other internet companies, such as broadband service providers." /14

We'll have to see what the rest of the doc says about the details there. 15/

On to some policy. Awkward: NTIA cites a litany of non-discrimination arguments ranging from the Postal Service Act to telegraphs and telephones to antitrust but omits #NetNeutrality regulations for ISPs, which the current FCC is in the process of trying to repeal. 16/

Lots of stuff about how platforms are now powerful and bad, collaborate with the Chinese government, a complaint that academics don't study social media bias (!), and how 230 is to blame. 17/

On to explanations of how modern platforms are different than CompuServe, Prodigy, and AOL. ( @jkosseff they don't seem to realize that CompuServe is still online!) 18/

Now to the fun part: legal arguments. 19/

First up, the Commission's power to regulate platforms. Holy guacamole: they're citing to Section 201(b), part of the (up until earlier today) reviled Title II of the Communications Act. Wuuuut? 20/

<many pages of explanation for why that's not as wacky as it sounds> 21/

<several more pages of history about 230 to distract from the wild authority argument> 22/

More history and a new argument that doesn't really get to the authority issue: courts misinterpreted Section 230 to apply to causes of action beyond defamation. 23/

Here's some more meat: "Ambiguities in Section 230." There are three on the menu. 24/

First, the NTIA tries to characterize the relationship between (c)(1) and (c)(2) as having some sort of ambiguity. Honestly, this is a nonsensical argument. 25/

NTIA says, without much explanation, that courts are applying (c)(1) without regard to (c)(2), saying that that renders (c)(2) meaningless. As predicted, it badly handwaves at some connection between the two that just doesn't exist in the text of the statute. 26/

This has something to do with ask #2, but I can't make sense of the proposed regulatory text after reading it about five times. Moving on. 27/ https://twitter.com/blakereid/status/1287876368470949889?s=20

Next up: alleged ambiguity in the "otherwise objectionable" prong of (c)(2). Lots of comments about how this ambiguity has caused problems in spam cases—cc @daphnehk 28/

Not quite clear where this is heading, but some old-school historical references to give context to what the NTIA believes "otherwise objectionable" refers to based on the preceding "obscene, lewd, lascivious, filthy, excessively violent, harassing" language. 29/

1909 amendments to the Comstock Act, the CDA itself ( @jkosseff you're gonna scream when you read this part), Pacifica, the CDA again, the statutory provisions about the V-Chip, the TCPA, telephone harassment statutes, etc. 30/

All this is a cover to dramatically narrow the scope of "otherwise objectionable," which the proposed reg is defined as "any material that is similar in type to obscene, lewd, lascivious, filthy, excessively violent, or harassing materials." 31/

Holy shit: NTIA doesn't even contend with the "whether or not such material is constitutionally protected" language. 32/

The idea here is that platforms can *only* moderate speech that is nominally illegal on various other statutes that govern other kinds of (mostly non-Internet) platforms. This would dramatically limit what platforms could do in conjunction with the (c)(1) <> (c)(2) ask. 33/

On to "good faith"—a citation to cases that are concerned with moderation for "anticompetitive purposes," but as expected, it's a cover for a much more sweeping change. 34/

The regulatory language is super long, but it has five requirements. 35/

First, the removal has to be consistent with terms of service, and those terms of service must "state plainly and with particularity [whatever the hell that means] the criteria the interactive computer service employs in its content-moderation practices." 36/

It's hard to keep the full context here, but this is a Rube Goldberg effort to somehow jam algorithmic accountability requirements in here. 37/

The second requirement is that the moderation falls in one of the categories preceding "otherwise objectionable" (see above). 38/

The third requirement is that the moderation doesn't happen on "deceptive or pretextual grounds [again, whatever the hell that means], and that the platform enforces its TOS consistently against "similarly situated [AGAIN, whatever the hell that means] material." 39/

The 4th requirement is badly written and mixes up ICS and ICP, but the platform must provide " timely notice describing with particularity the interactive computer service’s reasonable factual basis for the restriction of access and a meaningful opportunity to respond." 40/

NBD, just jamming a huge counternotice and process regime into Section 230. 41/

There is an exception for when the platform has "an objectively reasonable belief that the content is related to criminal activity or such notice would risk imminent physical harm to others." 42/

Oops, there are only (!) four requirements in the good faith reg, not five. Moving on. 43/

The third and final source of supposed ambiguity is the definition information content providers (ICPs). [insert "ICPs, how do they work" GIF here.] 44/

This gets to the ask about the "contribution" doctrine—when does a platform cross over from just hosting speech to effectively *becoming* the speaker? 45/

The reg would treat an ICS as an ICP when it "substantively contribut[es] to, modif[ies], alter[s], present[s] or prioritiz[es] with a reasonably discernible viewpoint, comment[s] upon, or editorializ[es] about content provided by another information content provider. 46/

I don't have a good feel for the contribution cases and am not sure what kind of an expansion this would constitute. (Help @daphnehk @jkosseff!) 47/

Oops: there's a fourth (!) alleged ambiguity: the meaning of being "treated as a publisher or speaker" under (c)(1). 48/

Whoo. This is basically just galaxy brain "platform vs. publisher" theory. ( @jkosseff, pour yourself several drinks for this part.) 49/

Oh man, this proposed reg is so badly written: "An interactive computer service is not being 'treated as the publisher or speaker of any information provided by another information content provider' when it actually publishes its own or third-party content." 50/

It's trying to carve out from "treatment as [a] publisher" acts of "actual publi[cation]." This is about the most confusing terminological choice ever made. 51/

Anyway, we get some definitions of what "actual publication"—where you, bear with me, are *not* being "treated as a publisher" for the purposes of 230 but *are* for purposes of some other law. 52/

Once circumstance: when an ICS "affirmatively solicits or selects to display information or content either manually by the interactive computer service’s personnel or through use of an algorithm or any similar tool pursuant to a reasonably discernible viewpoint or message." 53/

A carveout: "without having been prompted to, asked to, or searched for by the user." As far as I can tell, this means that platforms lose 230 protections if they do anything to sort posts without an affirmative request from a user. 54/

The second circumstance: if an ICS "reviews third-party content already displayed on the Internet and affirmatively vouches for, editorializes, recommends, or promotes such content to other Internet users on the basis of the content’s substance or messages." 55/

Cont'd: "This paragraph applies to a review conducted, and a recommendation made, either manually by the interactive computer service’s personnel or through use of an algorithm or any similar tool." 56/

Not sure entirely what the subtext is here, but seems like it's going after search engines? 57/

Anyway, we get some carveouts from the "actual publication" circumstances: 58/

"[P]roviding content in a form or manner that the user chooses, such as non-chronological order, explicit user preferences, or because a default setting of the service provides it, and the [ICS] fully informs the user of this default and allows its disabling" 59/

So you can set defaults up but the user can turn them off. 60/

"[T]ransmitting, displaying, or otherwise distributing such content, or merely by virtue of moderating third-party content consistent with a good faith application of its terms of service in force at the time content is first posted." 61/

So actually just posting the content or moderating it consistent with your TOS is copasetic. 62/

Now we're back to the FCC's authority, this time in the context of some kind of disclosure mandate. 63/

Interesting: NTIA leans hard into the notion that ICS are non-regulable information services. This normally doesn't precede an argument to regulate them. But! 64/

They propose following Section 163 and 257 of the '34 Act, more or less consistent with what the FCC did to avoid completely blowing up the transparency regs in the Open Internet repeal. 65/

Setting the authority aside, here's the disclosure mandate: 66/

"Any person providing an interactive computer service in a manner through a mass-market retail offering to the public shall publicly disclose..." 67/

(Not much here on the weirdness entailed in calling a freely offered service a "retail offering," but I digress. 68/

What must be disclosed? "[A]ccurate information regarding its content-management mechanisms as well as any other content moderation, promotion, and other curation practices of its interactive computer service..." 69/

That doesn't give a lot of specifics. The information must be "sufficient to enable (i) consumers to make informed choices regarding the purchase and use of such service..." 70/

(Hard to see how this is an enforceable standard. Also, reminder that most consumers don't "purchase" these services.) 71/

"...and (ii) entrepreneurs and other small businesses to develop, market, and maintain offerings by means of such service." (I guess we're gonna try to throw in app store regulation while we're at it?) 72/

Bonus: "Such disclosure shall be made via a publicly available, easily accessible website or through transmittal to the Commission." I guess the FCC is now going to keep a registry of every bulletin board on the Internet? 73/

Maybe the idea is that every ICS will host this disclosure on their own sites. Privacy lawyers, watch out, we have a fun new compliance-y thing that will have your clients asking if you also practice telecom law! 74/

Whew, that was a lot. If you want the TL;DR: version—the NTIA 230 petition would radically reshape the way just about any Internet platform with user generated content would have to operate, but... 75/

...the legal authority for the FCC to implement any of these asks is as flimsy as many telecom experts (including @haroldfeld, notably) predicted beforehand. The analysis of alleged ambiguities in Section 230 is just embarrassingly wishful thinking on NTIA's part. /fin