I wouldn't put it that way. The problem is that a ton of the stuff impinging on us is not knowledge. It's propaganda, fluff, stupid memes, click bait, and systems with interfaces designed to be addictive and to act like Skinner boxes.

What I really see happening is a "part 2" repeat of the spam apocalypse that hit e-mail, Usenet, and other federated systems on the early Internet. In the late 1990s spam destroyed most of these systems. E-mail was so valuable it had to be saved, but saving it has mostly involved delegating it to a small oligopoly of vertically integrated e-mail providers with the resources to fight spam. Usenet and other less well known federated systems were destroyed.

I think we are now seeing a similar kind of spam assault against the open web, social media, and the news. The latter with the whole "fake news" phenomenon really resembles a Sybil attack. In recent years I've seen an avalanche of totally fake news sites that look like legitimate newspapers popped up (apparently from templates) and used to spam social media with a whole slew of unfounded allegations and other assorted bullshit. The role they've had in the election has gained a lot of press but it's in no way limited to that or even to politics. A lot of it seems to be shilling for quack medical products and other traditional mainstays of... spam!

I've started calling this whole phenomenon "spam 2.0." I'd define spam 2.0 as the semi-automated bulk creation of fake news outlets, fake social media profiles, and hoaxes and their use to engage in the coordinated manipulation of opinion.

I'm also anticipating the rise of spam 3.0. Spam 3.0 will be fully automated and AI-powered, with AI being used to create an individual "persuasive agent" for each human target. Think of it as massively parallel individual con artistry and/or "spear phishing" at scale. Social media companies are already pioneering this tech, but the spam apocalypse will really hit when it goes public and you can do it with readily available tools.

I find it rather depressing but for a while I've been predicting that the first use of AI that can (even borderline) pass the Turing test will be con artistry, cult indoctrination techniques, and demagoguery at scale. It's really pretty scary to the point that I'm wondering if the use of AI to persuade humans should be outlawed. Of course then you get the "then only criminals will have guns" effect. This is particularly true on the borderless Internet where it's unlikely to be outlawed everywhere. Users of this tech would just go offshore.

An alternative to outlawing might be to disseminate the tools as widely as possible to eliminate any comparative advantage and render the whole thing moot. If this destroys social media in the process then so be it. Going to happen anyway.

Edit: the bottom line is that "cognitive self-defense" is going to become a thing. These "warning signs for tomorrow" are becoming relevant:

http://www.aleph.se/andart/archives/images/warning4.png

We are living in a cyberpunk novel.