Comments on the PROMISE Act
Summary of the Bill
The bill would require interactive computer services (ICSes) that engage in interstate commerce, including “common carriers,” to publish an “information moderation policy” (that meets some statutory specifications) about how they moderate content. The failure to publish a moderation policy, or any noncompliance with the announced policy, would violate Section 5 of the FTC Act and would be enforceable by the FTC. There would be “a rebuttable presumption that a deceptive policy statement is material and likely to cause injury.”
What’s Good About the Bill
Four strengths of the bill:
1) It only addresses one problem. This is far better than omnibus reform bills like the SAFE TECH Act or the PACT Act, which package together a mélange of bad ideas into a tangled and complicated mess. If Congress wants to have a serious conversation about Section 230, single-reform bills are the way to do it.
2) It doesn’t touch Section 230 directly. Though the bill clearly targets concerns about Section 230, it doesn’t create any new Section 230 exceptions/carveouts or condition the availability of Section 230’s protection on new obligations.
3) It designates the FTC as the sole enforcement agency. It makes sense to treat transparency obligations as a consumer protection question and to designate the FTC–our nation’s top consumer protection expert–as the enforcer.
4) Transparency is a comparatively light-touch approach. Of the various themes for Section 230 reform, I think transparency is the regulatory vector that would do the least amount of harm.
Some Problems With the Bill
Even though the bill has some positives, it’s not good overall. Some of its problems:
1) Does It Solve Any Problem? The bill would require ICSes to disclose the categories of content they moderate/ban, the processes they use to moderate content, and if/how they notify users about actions taken, reasons for the actions, and appeal options. All of these disclosures are supposed to be disclosed “in a publicly available and easily accessible manner” and “in plain, easy to understand language.” As we know, the lawyers are going to make mincemeat out of disclosure requirements like this. The disclosures will be written to minimize liability, not enlighten consumers, so they will be drafted with vague and inscrutable language and filled with caveats and disclaimers. The bill says it seeks to help “users to make informed choices,” but any lawyer-drafted ICS policies won’t advance that goal.
A reminder to bill drafters: there is a robust literature on best practices for designing mandatory disclosure schemes. Start with Archon Fung et al’s Full Disclosure book.
2) Can It Be Operationalized? Let’s assume ICSes, despite their self-interest in laconic disclosures, actually did what the bill drafters want and provide useful details about their policies. The bill does not address how to handle several routine circumstances:
- What if the ICS encounters an edge case where the published policy would not yield the right result? Every content moderation policy inevitably encounters new facts that weren’t contemplated by the drafters, and services typically handle them judiciously and then learn from the exception. Under the bill, the service would be forced into the wrong result.
- What if the ICS encounters a problematic new set of facts that the published policy does not address? I believe the bill would prevent the ICS from taking any action (at least until it adopts a new policy).
- How can an ICS amend its existing published policy? The bill doesn’t address amendments at all; and FTC jurisprudence may restrict the ability of ICSes to change the deal for existing customers who signed up under different promises. If so, the bill would freeze published policies, possibly forever. That can’t be the right outcome.
Per point #1, to ameliorate these concerns, lawyers will draft vague and capacious policies.
3) Is It Constitutional? The bill expressly governs publication functions. The definition of “moderate” covers removing/restricting access to third-party content, “editing” or otherwise altering third-party content, and adding a warning/fact-check/other label to third-party content–all clearly editorial activities. So this bill, like the PACT Act, raises the question of whether legislatures can (1) mandate disclosures of publishers’ editorial practices, and (2) punish them for exercising that editorial discretion in new or different ways. In my blog post on the PACT Act, I gave the analogy to a book publisher receiving unsolicited manuscripts for possible publication. Does the First Amendment allow a legislature to mandate that the book publisher articulate its editorial standards and how it does its editorial work? See the McManus case. My next big paper will address the potential constitutional limits the ability of regulators to review and second-guess editorial decisions.
4) Costs Borne by Small UGC Services. I am generally skeptical of regulatory attempts to distinguish between larger and smaller Internet services. However, they are helpful when a bill imposes costs that will jeopardize smaller services. In this case, the bill would treat small UGC services (say, like my blog comments) the same as Facebook. The bill would require me to make disclosures about my moderation policies that I don’t currently make and don’t really want to spend the time or money to do. I would turn off blog comments rather than try to comply with the law. Many other smaller UGC services would make that choice too. So even a purportedly light-touch bill will likely shut down parts of the Internet.
Prior Blog Posts on the 117th Congress’ Efforts to Destroy Section 230