From Judge Ruth Bermudez Montenegro (S.D. Cal.) Oct. 22 in Jalisi v. City of Poway:
The City of Poway operates a publicly accessible Facebook presence intended to share updates that may interest residents and the broader public. This page allows comments from anyone, and members of the public routinely respond to posts there. The plaintiff, a Poway resident, has for years voiced criticisms of various city decisions by speaking at council meetings, writing to City officials, and posting pointed remarks on the City’s Facebook page. City officials and staff involved with managing the page appear to view Plaintiff as a critic.
Over the last year, Plaintiff’s comments on the City’s Facebook page have been repeatedly “hidden” or “shadow banned”—not deleted, but concealed from the view of other users even though they remain visible to Plaintiff when he checks the posts. The City’s Social Media Policy states that the City’s social media presence operates as a limited public forum and that content will be monitored and moderated. The Policy explains that user-generated posts will be removed if they meet criteria such as: “abusive, vulgar, offensive, profane, sexual, threatening, deliberately untruthful or defamatory language or content.” It also notes that postings from a public user account that contain inappropriate or disruptive content, or are believed to be spam, may result in the public user account being blocked from any further activity on the City’s social media account at the discretion of the City Manager’s Office.
The Plaintiff contends that the City is masking his comments to suppress his critical views. The City states that none of its employees have hidden or removed any of Plaintiff’s comments; instead, Meta’s Moderation Assist Program has hidden them because it tagged Plaintiff’s Facebook profile as a non-human “bot” account. Moderation Assist is described as necessary because the City does not possess the resources or ability to monitor every single comment in real time and to manually hide bot comments, especially given that these comments can be submitted 24/7. The City contends that Plaintiff’s remarks were hidden because Moderation Assist flagged him as a bot: “Moderation Assist will automatically hide a post if (1) the commenter has no friends or followers, (2) the post contains a link, or (3) the commenter lacks a profile picture.” Plaintiff does not have a profile picture, has only one Facebook friend, and has no posts on his own profile, which “are typically strong indicators that the account is a non-human bot.”
The city agreed that moderation based on viewpoint would be unconstitutional, and agreed to “manually unhiding Plaintiff’s prior Facebook posts.” But it opposed Plaintiff’s request to “remove any restriction that causes Plaintiff’s comments on the City Page to be hidden from the public”; it argued that it needed to continue using Moderation Assist and could not disable it for Jalisi alone. The court therefore ruled in the City’s favor on this issue:
[T]he Court must determine whether the City’s Facebook page operates as a limited public forum or a designated public forum. Even when evaluating laws regulating expression in quintessential public forums, courts first consider whether the law is content-based or content-neutral. …
At the outset, given the limited and preliminary record before it, the Court notes it is likely that Plaintiff’s comments were hidden not by a City policy, but by Meta’s Moderation Assist Program. But even if the restrictive effects of Moderation Assist could be attributed to the City, the Court does not find the restrictions unconstitutional.
First, the restriction is facially content-neutral. According to the Carmichael Declaration, profiles are not flagged as bots by looking at the content of the message; rather, they are flagged based on predictive factors such as the absence of a profile picture or the lack of other Facebook friends. Consequently, determining whether a comment may be hidden has nothing to do with the substance of the comment.
Second, the restriction can be justified without reference to the content of the regulated speech, and there is no indication that the restriction was adopted by the government due to disagreement with the message conveyed. The City uses Moderation Assist because it lacks the resources or capability to monitor every comment in real time to manually hide bot comments. There is nothing in the record suggesting the City adopted Moderation Assist out of opposition to any particular expression. Indeed, Plaintiff’s evidence that identical comments published by other users remained visible strengthens the finding that the restriction is content-neutral; critical speech is clearly tolerated by the City. Therefore, the Court concludes that the restriction is content-neutral and subject to intermediate scrutiny.
To survive intermediate scrutiny, the restriction must “further an important or substantial governmental interest,” that interest must be “unrelated to the suppression of free expression,” and the incidental restriction on First Amendment freedoms must be no greater than necessary to advance that interest.
The City contends that its interest in employing Moderation Assist is to “prevent non-human bot spam postings on its Facebook page, which can overwhelm public pages and make it much harder for the public to obtain important information and participate in the conversation.” The Court finds that the two-way exchange of information between citizens and the City constitutes a substantial governmental interest.
Moreover, that interest is “unrelated to the suppression of free expression.” In fact, the interest runs counter to suppression of speech; it seeks to ensure that residents have meaningful access to the exchange of ideas with their neighbors and their government. The interest concerns limiting the influence of non-human bots, but bots themselves are not “human beings with First Amendment rights” and do not constitute corporations with independent First Amendment rights.
Finally, although the record shows there are “incidental restrictions” on speech, such as hiding comments posted by humans flagged as bots, these incidental restrictions are “no greater than necessary to further” the governmental interest in question. “A key aspect of this inquiry turns on whether the government seriously attempted to address the problems it faces with less intrusive tools readily available to it.” Based on the early record, the Court finds that the City has no viable alternative methods to prevent non-human bot accounts from flooding its Facebook page. This is not a case where there is an obvious, alternative regulatory approach that would achieve the government interest without censoring any speech….
Alan B. Fenstermacher and Jayson A. Parsons (Rutan & Tucker) represent the City.
