Moderating Amazon Web Services

I’m trying something a little different here: a forum-style post where I pose a question or problem, usually after describing it at some length, and ask for your comments and feedback. To be honest, I’m stealing this format for my own purposes from Helena Fitzgerald’s excellent newsletter griefbacon, but where her newsletter (including subscriber contributions) is personal and literary, this one is embarrassingly analytic. So I’m fishing a little bit here, and hoping to pull up something good.

Since January, I’ve been working on a story about Amazon Web Services. It was motivated by AWS booting the social media network Parler off its platform for violating the company’s terms and services. Since then, AWS has done the same for accounts linked to the NSO Group, which was linked to surveillance software that was used to spy on journalism organizations worldwide. (AWS’s response came after an Amnesty International report claimed that NSO had recently switched to Amazon’s CloudFront service.)

The problem here is that there are no clear guidelines as to what, if any, services a backbone internet infrastructure provider like AWS should refuse to provide to a client. While public-facing platforms like YouTube or Facebook might act relatively quickly (in some cases automatically, and in some cases using software like that provided by AWS) to pull down a video that violates copyright or is deemed abusive, commercial platforms like AWS are rarely called upon to stop supporting entire platforms. Yet in other cases this lack of accountability and the absence of oversight is exactly what allows truly atrocious behavior to continue.

Another distinction one might make is between Amazon as a public-facing storefront and again, as an infrastructure provider for other, perhaps competing applications. Amazon’s Kindle store (or print bookstore) might come under public pressure for profiting heavily off controversial or misleading books, but that seems different in some degree from a case where Amazon is providing hosting services for a separate bookstore selling the same book, or an app or video streaming service offering that same content. Should Amazon be in the business of policing everything its wholesale customers do? If so, how would that work?

You could also take a maximalist position, in either direction, offering close scrutiny or virtually total freedom. Amazon is, to some degree, either responsible (or in no way responsible) for anything done on its platforms, using its hosting or other cloud services, or sold using its marketplaces. Any distinction one could make between these layers is formal or fictitious, and Amazon should be equally accountable (or equally free) to offer anything.

So I put the question to you, the readers, whom I consider by and large a pretty knowledgable bunch: in what cases should Amazon, or any other commercial or internet infrastructure provider, intervene when its own services and technologies are being used for something objectionable or unlawful? Should the laws protecing Amazon from liability for third-party use of its platforms be changed? Or are there standards different from legal liability which Amazon should observe to preserve its own brand identity and legitimacy?

It’s a legal problem, a technical one, and a practical set of thorns all rolled into one. I’m talking to technical and ethical experts, and I’m talking to you: what do you think should be done?

This email, including its web version, is restricted to paying subscribers for the Amazon Chronicles. You can either reply directly to me (in which case I’m the only one who’ll see it), or post a comment to help continue the discussion. Please be thoughtful and charitable in your responses to each other.