This new EU legislation granting providers the right to “voluntarily” scan private messages doesn’t break encryption, or take us to a regime of mandatory mass surveillance. But it definitely sets the stage.
https://twitter.com/japi999/status/1412667455319728131
What’s remarkable about this stuff is that it’s phrased as “protecting children from child abuse”. And as a parent I appreciate that. But has anyone explored, empirically, if any of this surveillance actually works to stop the problem?
Here in the US we’ve built an enormous surveillance system to detect instances of child sexual abuse material, it’s been running for years, and the number of reports is going up exponentially.
How many pedophiles are there? Isn’t it a smallish number?
How many pedophiles are there? Isn’t it a smallish number?
It’s remarkable that we’re all so eager to build a surveillance system that has basically unlimited machine access to all of our private messages, and nobody seems at all interested in whether any of this stuff works to deter the underlying crimes.
Building systems to detect CSAM is hard. Detecting “grooming” behavior is even more invasive, since it requires machine learning techniques that understand a wide variety of interactive human communication.
Building the infrastructure to surveil communications at this level of detail (and committing to keep it operating as more and more services inevitably add end-to-end encryption) is not a business that democracies should be undertaking.
There is functionally no difference between a machine system that can detect “grooming behavior” and one that can detect literally any other human communication pattern.
Once you hold open the door for providers to solve this problem, you’ve committed to supporting automated surveillance with capabilities that could make the Chinese government envious. It’s only a question of how you train them. //
• • •
Missing some Tweet in this thread? You can try to
force a refresh
