Thread by @doctorow

Last year, the EU introduced the #DigitalServicesAct as part of a package of anti-monopoly measures aimed at US-based Big Tech. EU regulators could lead here, since - unlike US regulators - they don't have to worry about Big Tech as a source of national "soft power." 1/

If you'd like an unrolled version of this thread to read or share, here's a link to it on pluralistic.net, my surveillance-free, ad-free, tracker-free blog:

pluralistic.net/2021/10/28/cli…

That's why the EU gave us the #GDPR, a serious and far-reaching privacy law. 2/

Unfortunately, the GDPR has been hamstrung by its enforcement problems. Under the GDPR, enforcement is delegated to each country's Information Commissioner's Office (ICO). 3/

The US Big Tech companies all fly Irish flags of convenience, and in return the Irish state bends over backwards to help them. It's not just maintaining the fiction that Big Tech's billions are floating in untaxable, international waters somewhere in the Irish Sea, either. 4/

As far as I can tell, the Irish ICO doesn't even put on pants in the morning. They just sit around in their pajamas, eating breakfast cereal and watching cartoons while the enforcement actions against GAFAM pile up on their doorstep.

pluralistic.net/2021/06/16/ins… 5/

With the #DSA, the EU had a chance to take another hack at reducing US Big Tech's corrupting, destructive conduct. 6/

The first draft was actually pretty fantastic, creating fairness and transparency obligations that would have really hit Big Tech where it hurt:

eff.org/deeplinks/2021… 7/

But in the months since, the process has been hijacked by the same EU bad-idea-havers that gave us the 2019 Copyright Directive. These people suffer from incredible brain-worms. 8/

They are convinced that they can fight bad content by forcing Big Tech to build algorithmic filters that will magically detect and block illegal speech.

This is a fantasy, and a dangerous one. 9/

You can't fix Europe's technological problems by subjecting every European's online life to the opaque, instantaneous judgments of American companies' black-box "AI" filters.

eff.org/deeplinks/2021… 10/

These systems make speech disappear in an instant, and then send the censored to the back of an appeals queue that can stretch for weeks, or months, or years:

pluralistic.net/2020/05/17/che…

They're not cheap, either! 11/

Google's Content ID filter - falsely billed as keeping Youtube free of copyright infringement - has cost $100m so far, and the bill is still mounting. 12/

Putting an obligation on all tech companies to build these filters will wipe out the entire EU tech sector and leave five or six US companies to rule Europe's digital future.

Apologists for the latest DSA proposals will say they're not requiring filters per se. 13/

They're just ordering tech companies to block all illegal speech, and any speech that resembles previously identified illegal speech. 14/

That's the same argument they made about the Copyright Directive, only admitting that filters were always the plan after the vote went through.

These new proposals will hold companies liable for their users' speech if they act as "active platforms." 15/

An "active platform" is one that categorizes, sorts, analyzes or otherwise manages its users' communication. Punishing companies for being "active" is no way to fight illegal content. Finding and removing illegal content REQUIRES that you first analyze and categorize it! 16/

All of this runs counter to EU law and principle. The GDPR, for all its flaws, bans subjecting people to "automated decision-making" without their consent. 17/

The foundational EU internet rules, the E-Commerce Directive, requires all content-removal rules to comply with "observance of the principle of freedom of expression." 18/

Europe can still save the DSA and with it, its citizens' free expression rights, and its nations' digital autonomy from US platforms. The EU could balance content removal with speech and privacy rights. 19/

It could make the appeals process for the censored co-equal with the process by which content is removed. It could give platforms a reasonable timeframe to assess takedown demands. It could ban algorthmic filters, rather than requiring them. 20/

It could nurture - rather than exterminate - small, European platforms, by requiring big US companies to interoperate with them.

If it fails, the risk goes beyond Europe. 21/

We've seen how bad internet laws reproduce themselves, and especially how autocratic states use these rules to justify their own internet censorship and surveillance regimes. 22/

If Europe legitimizes the filternet, it will be a signal to the world's tyrants that they should try their hands at it, too. 23/