Talk Notes on the Constitutionality of Transparency Requirements for Internet Services
I’m writing about laws that mandate the transparency of editorial activities by Internet services. In this context, those mandates can take many forms, including mandatory disclosures of editorial policies, mandatory explanations of decisions, and publication of statistics about editorial operations (what is often called “transparency reports”).
Transparency obligations have come into vogue as a regulatory solution because:
- they bypass partisan gridlock, where Democrats want more content removals and Republicans want less content removals. Transparency is a “third way” that both parties might agree upon.
- in theory, transparency helps the marketplace hold Internet services more accountable.
- most likely, the goal of mandatory transparency is to control content indirectly and produce information that will fuel future lawsuits.
Numerous federal bills have proposed mandatory transparency obligations, such as the PACT Act. I’ll emphasize these two state laws requiring transparency:
- Florida requires Internet services to publish their editorial standards and provide mandatory explanations of some editorial decisions.
- Texas requires all of that and more: Internet services must “disclose accurate information about content management, data management, and business practices”; and they must publish transparency reports with exacting details
Everyone seems to love transparency as a policy solution, but I am possibly throwing cold water on the enthusiasm. Mandatory transparency may be a policy dead-end because they likely conflict with the First Amendment.
The core problem: for governments and plaintiffs to doublecheck the accuracy of any disclosures, they will necessarily intrude into the editorial operations of Internet services. To the extent that Internet services are making editorial decisions like other publishers, such intrusions aren’t permitted.
[Note: if you believe that Internet services aren’t engaging in editorial activity, the rest of my paper won’t persuade you.]
Examples of how courts have identified limits of mandatory transparency obligations:
- Herbert v. Lando (Supreme Court 1979): the plaintiff was allowed to depose the editor about his editorial processes in defamation lawsuits where “actual malice” was at issue, but the Supreme Court rejected other information-gathering to satisfy curiosity or for a general purpose like the “public interest.” The prospective disclosures required by Florida and Texas are exactly that–designed to satisfy curiosity or advance the “public interest.”
- Washington Post v. McManus (4th Cir. 2019): the Fourth Circuit struck down (1) a requirement that Internet services disclose information about political ad buyers, and (2) an obligation to collect record and let a government agency inspect them. The court says this inspection right “brings the state into an unhealthy entanglement with news outlets.”
These concerns about intrusion into editorial decision-making distinguish mandatory transparency obligations imposed on Internet services from the many other legitimate and constitutionally authorized mandatory disclosures for marketplace goods and services, such as nutrition labels or country-of-origin disclosures. Similarly, restyling mandatory disclosure laws as standard UDAP/consumer protection doesn’t address the underlying entanglement problems.
Mandatory disclosure laws can also raise concerns about compelled speech and user privacy.
If the First Amendment limits the ability of governments/plaintiffs to obtain discovery, then any disclosures by Internet services are probably fiction/propaganda. Problematically, this concern applies to voluntary disclosures too.
Possible workarounds to the constitutional limits:
- require Internet services to get mandatory third-party audits, just like publicly listed companies must get audited financial statements.
- mandate information governance procedures, e.g., a Chief Transparency Officer with statutory duties.
- facilitate the work of independent researchers, including liberalized data scraping.
My paper will explain why these workarounds may be policy dead-ends too.
__
I have a rough draft of the paper that would benefit from comments. If you are interested in reviewing it and sending me comments, please email me. Otherwise, I expect to have a public draft in about two months.
Want more? See my RightsCon talk on this topic.