It's not the post office

I've mentioned many times previously that social media requires regulation. I've explained my position regarding the ideals of free speech often enough that I don't feel the need to do that again in any detail. Suffice to say that the 'marketplace of ideas' metaphor is if anything too successful : real markets are full of people buying and selling utter crap. By itself, the market is just not a valid solution for optimising anything very much; with proper oversight, it's a different story. Nobody seriously believe in a totally free market. Nobody.

Today I just want to briefly cover a different angle from a recent discussion. Sometimes a counter-argument to regulation is made by way of comparison to more traditional media : TV, radio, telephone, the postal service, etc. We don't routinely eavesdrop on phone calls (at least, we're not supposed to), so why should social media be different ?

My previous answer has been that social media is just not fairly comparable to traditional media. I stand by this assertion : it offers mixed media, audience reach and participation, and longevity of content in a way so unique that it deserves to be treated as its own format. To compare it to a phone call is a bit like comparing a train to a horse : yes, you can use both of them to get around, but they're really not the same thing at all - even from just a transport perspective.

But it's not really very satisfactory to just point out that social media is different. After all, if we were asked why trains were important in the Industrial Revolution, we could hardly be content by saying "because they were better than horses". Far better to define the exact differences, the relevant rules that apply in all conditions to decide when regulation of content is legitimate. Otherwise we risk it becoming a special case, with the alleged need for monitoring being little better than "because I want to". If we can state the general rules, however, then no-one can accuse us of unfairness. 

And anyway, only pointing out that social media is different to other media is like saying that it's only the consequences which matter, i.e. that the ends justify the means. That's something I'd rather avoid.

Andreas Geisler's (yes, him again) suggestion in the discussion is that it's about promotion. Television and radio are clearly publishers, but the telephone and postal services are merely delivery agencies. They explicitly have no knowledge of the content they convey, or even anything of the sender and receiver besides names and addresses. This lack of knowledge is an explicit feature of the service, why we largely feel able to trust such services with confidential information. Without it, these services would be unusable. 

Additionally, but secondarily to this, they don't host content either. They don't store anything a minute longer than they need to deliver it. They don't make the content available to anyone besides the intended recipients. True, they don't actively prevent resharing content either, but their very nature makes this difficult enough to deter almost everyone besides the truly obsessed. The responsibility for what happens to the content they deliver lies entirely with the recipient : they themselves accept responsibility for delivery only - nothing else.

Compare this with television, radio, or book publishers. Those both generate, host, and disseminate content, and as such are definitely responsible for the content they distribute. As such they are subject to very different expectations and regulations.

What about social media ? Well, knowledge of the content posted is intrinsic to the business model, as is providing tools for resharing. You create the content, yes, but they facilitate both its delivery and availability. In stark contrast, the phone service would very much like you to talk to lots of people, and the postal service would be a lot happier if everyone started writing letters again, but they don't provide any direct means to keep you engaged. You don't want your phone company automatically interjecting when you're about to hang up with a voice saying, "did you also want to talk about the geopolitical situation in North Korea ?". It'd be like if the Matrix was real but run by Clippy.

Likewise if you hire a delivery van, they don't go out of their way to find you more stuff to move from point A to point B. That's just not what delivery companies are for. Nobody would want them to be like that, because it wouldn't work.

(There are plenty of specialist exceptions, e.g. delivering medical products, but we can probably safely ignore these here.)

And that basically is what social media outlets typically do. They actively recommend you particular communities, people, topics or other ways to find content you're interested in. Not even the worst cold callers will phone you up and give you a list of ten other people interested in donkeys because you happened to go to a farm zoo one morning, but that's quite a reasonable analogy for how social media works. What would be unbearable for the delivery sector is not merely desirable but to a large degree actually essential in the social media arena.

Incidentally, note how weird social media is compared to other outlets. TV is definitely a publisher, but it doesn't encourage resharing content. It likes you to promote references, yes - telling your friends about a good show boosts ratings, and TV likes that. But that's not sharing the content itself at all, any more than sharing the address of a museum is equivalent to giving everyone a physical copy of Tutankhamun's death mask. Nor does the phone service make it easy to record and reshare. One-off, user-initiated delivery is not just comparable to prolonged hosting and sharing within a wide community.

These basics hosting and sharing features are common to all the major social media outlets. And it's this intrinsic awareness of content which, it's suggested, is what makes them publishers. A choice is made based on content to form links between people. Indeed, creating and promoting a recommendation is itself a form of content creation and publishing.

And it's also a tacit endorsement of the content itself, if only to a limited degree. Just as a publisher authorising a cookbook doesn't have to think every meal is delicious but does have to ensure it won't actually poison anyone, so does a social media company recommending white supremacist groups to non-members bear some measure of responsibility for this. They're not wholly responsible for the content, but they can't be given carte blanche either. They didn't create the original content, but they do create recommendation links - which are essentially derivative works.

Crucially, this is by choice. No-one is forcing social media to exist, nor for it to recommend anything to anyone. Instant messaging services don't do this - they can still be fairly called delivery services. But in choosing to host, share, and recommend, social media companies make knowledge of content fundamental to their success. It's this aspect of choice which means that even if this is done algorithmically, their responsibility for this is at most only slightly diminished. This makes them publishers. Perhaps they are not 100% responsible for all content published on their site, but they are absolutely, 100% responsible for all content generated by themselves, even algorithmically.

To be fair, not all of it works the same way. There's no reason you can't have a system where the user alone is responsible for what they sign up to. You could in principle have a system where there are absolutely no recommendations whatever. Would this still be a publisher, or would it instead be reverting to a delivery agency ?

A better analogy here might be a bookshop. If the shop decides to label and organise the books to help with sales, they are responsible for the choices they make. They too are generating content (however limited) and making it available, i.e. they are themselves publishers. Even this can carry consequences, such as putting the Bible in the "history" section instead of the "religious studies" area.

But if the bookshop decides not to do this, to have a great big heap of books and allow anyone at all to come in and buy anything... that's probably not a publisher. It's also extremely stupid, like selling random items from an uncatalogued storage locker. What's this, an automatic rifle ? Sure ! A smallpox sample ? That'll be £3.50 sir, there you are... okay, such a shop would no longer be a publisher, but it would also be stupid and unworkable.

Yet the social media outlets I use are actually somewhat close to that. They don't provide any recommendation systems themselves, but let users set their own hashtags to label their own content for others to find. Needless to say, this is open to abuse, and using the wrong hashtags is one of the major annoyances. But actually it mostly works.

But does it alleviate the responsibility of the host ? At least partially this is surely the case. They did not create the hashtags the users generate, so cannot be directly or entirely culpable for that : they aren't publishing anything. If they have no control over who joins the service, then we might even extend this to a total mitigation of responsibility - just as a phone company cannot be held responsible if a terrorist uses a public phone, but could be if there was an order barring them from having a private telephone but they allowed an installation.

So there are still no good direct analogues for social media. But there are at least a variety of comparable situations we can use for broad guidance, which I think goes quite some way towards settling the question. Anything you generate and make available counts as publication, hence a service which recommends contacts is a publisher. If you don't, and don't set restrictions on who can join, you're probably closer to being a delivery agency.

The model of a purely user-driven service is a grey area. The wider question, of course, is whether this is a good idea. While we should have means-driven, not ends-drive, rules, we still should consider the consequences. This power to infinitely generate and duplicate multimedia content, to share it globally with strangers, to directly interact with people in prolonged, recorded conversations, is unprecedented. True, for now mainstream media still dominates most people's thinking. But we would be fools to assume this will last indefinitely, or to think that the regulations which we accept for more familiar delivery services are perfect analogies. Evading the label of "publisher" somehow feels hollow if it still allows for the viral spread of misinformation.

For now, the only conclusion here is that yes, social media services are indeed largely publishers, and at a minimum must accept responsibility for anything they themselves create in relation to user's own content. How much further we should take this I leave for now as an open question.