The end of Roe v. Wade is leaving web platforms struggling to handle moderation — and data safety — questions around newly instituted abortion bans.

For years, web platforms have faced relatively little backlash for pledging to remove posts that violate US law, even if there’s debate over whether specific content meets that standard. Policies that “match the laws of the country” — Elon Musk’s initial description of his plans for Twitter — are typically code for very permissive moderation in the US, where the First Amendment has historically protected a huge swath of content. The 2018 FOSTA-SESTA carveout, a rare exception, targeted the marginalized community of sex workers and used the justification that it was fighting the rightfully loathed cause of coercive sex trafficking.

But the present and coming state laws against abortion may change that. The laws — currently active in several states — ban a medical procedure that most Americans believe should be legal, and some groups want this ban to cover even information about the procedure. Large parts of the web have mobilized in response, offering to put up people seeking abortions and ship abortion pills, publicizing information about self-managed abortion, and (in rare cases) calling for violent resistance.

That’s already raising questions for Facebook and Instagram’s parent company, Meta. Earlier this week, both platforms removed posts that offered to mail abortion pills, citing rules against offering to “buy, sell, trade, gift, request, or donate pharmaceuticals.” According to a report from The Intercept, it also banned praising or supporting an abortion rights protest group called Jane’s Revenge, which has claimed responsibility for setting a fire at the office of an anti-abortion group, among other incidents.

These decisions aren’t particularly surprising. Shipping abortion pills to states like Texas could run afoul of those states’ laws, and Meta has bans on selling drugs like cannabis as well. Jane’s Revenge supports a cause many people find sympathetic, and the extent of the group’s size or activity is far from clear, but it’s claimed credit for crimes and called for a “night of rage” and “drastic measures” against anti-abortion infrastructure.

That said, critics have questioned the level of scrutiny abortion-related content has gotten. Instagram briefly removed the account of an abortion services locator and limited the hashtag for “mifepristone” on the grounds that it was being used in posts that violated guidelines. Jane’s Revenge was reportedly designated a “Tier 1” priority group, a higher restriction than the Oath Keepers and Three Percenters insurrectionist groups.

These decisions could eventually make their way to the Oversight Board, which theoretically exists to make difficult, nuanced moderation calls with some independence from Meta. But, with Roe v. Wade overturned barely a week ago, the Oversight Board hasn’t taken up any abortion-related cases. (A spokesperson acknowledged an email from The Verge asking whether it was considering the issues, but the organization didn’t respond by press time.) It seems plausible it will do so in the future — possibly drawing lines between things like providing medical information and directly facilitating a procedure.

But, even if it takes on the issue, that wouldn’t stop states and politicians from going after platforms that host abortion-related content. A model anti-abortion law from the anti-abortion National Right to Life Committee would ban offering information about getting an abortion illegally, and South Carolina lawmakers have already proposed an apparent version of it in the state legislature. Section 230 of the Communications Decency Act helps protect sites from liability right now, but it’s facing numerous challenges in courts and Congress.

Abortion isn’t the only issue where platforms could end up facing a patchwork of requests to take down content. As Politico notes, 34 states have introduced (and two have passed) bills governing social network moderation; many require sites to leave specific content up, but others demand takedowns of things like medical misinformation. Meanwhile, there’s a growing political movement to stop booksellers and libraries from letting minors access LGBTQ-related books, and if it continues to escalate, the fight could easily move online.

Web platforms have always had to deal with following laws their founders don’t necessarily agree with, and they’ve often acquiesced. Twitter limited hate speech in European countries while aiming to serve as “the free speech wing of the free speech party,” and Amazon recently restricted LGBTQ searches in the United Arab Emirates, among many other examples. But they’ve also been called upon to draw moral lines that go beyond following the laws of a given country. Apple, for instance, has been excoriated for following Chinese demands for censorship and access to user data. In the US, sites that submit to abortion-related police requests could face the same censure.

Companies like Meta have practical incentives to avoid being seen as partisan inside the US, and they’ve taken pains to do so, even when it involves deliberately putting their thumb on the moderation scale. While many have expressed support for employees’ reproductive choices, they’ve been quieter about any broader stance on the end of Roe v. Wade. But as abortion bans spread, that’s going to become more and more difficult — and eventually, they’ll have to make a call.