Tomorrow, the European Parliament’s Committee on Civil Liberties, Justice and Home Affairs (LIBE) will vote on a legislative proposal that would allow e-mail, messaging and chat providers to automatically search all personal electronic mail and messages of each citizen for presumed child pornography or child grooming (so-called chat control or e-privacy derogation). On 29 April, representatives of the EU Council and Parliament struck a deal behind closed doors. If the search algorithms flag a message, users will be reported to the police and their communications disclosed. The complete text of the proposed regulation is available as of today. With the exception of the Greens/European Free Alliance and Left groups, a broad majority is expected to vote in favour.
Shadow rapporteur Patrick Breyer (Pirate Party, Greens/EFA group) warns:
„This regulation means all of our private e-mails and messages will be subjected to privatised real-time mass surveillance using error-prone incrimination machines inflicting devastating collateral damage on users, children and victims alike. Countless innocent citizens will come under false suspicion of having committed a crime, under-agers will see self-generated nudes (sexting) fall in wrong hands, abuse victims will lose safe channels for counselling. This agreement sets a terrible precedent and is a result of a campaign of disinformation and moral blackmailing from foreign actors. Unleashing such denunciation machines is ineffective, illegal and irresponsible. It’s using totalitarian methods in a democracy.”
Breyer points to a legal assessment by a former Court of Justice judge according to which indiscriminate chat control violates fundamental rights. A poll published in April also shows that 72% of Europeans oppose indiscriminate screening of their private messages.
The political agreement was possible after a majority of Parliament negotiators gave up most of the few limitations and safeguards the Civil Liberties Committee had requested last year:
- language stating that technologies “should not include systematic filtering and scanning of communications containing text” has been deleted
- using error-prone “artificial intelligence” to search unknown images and videos for possibly illegal content will be permitted
- there will be no maximum error rate applicable to search algorithms, nor will “artificial intelligence” systems require prior authorization by data protection authorities
- correspondence with doctors, psychologists, lawyers etc. covered by professional secrecy will not be exempted from monitoring
- despite error rates of up to 86% according to police statistics, providers may automatically report “known material” to the police without human verification
- police will not be required publish statistics on the number of false reports and the number of users prosecuted as a result of chat control
- providers will not need to notify users they report to the police
The Commission is preparing follow-up legislation to make chat control mandatory for all providers, potentially including services that are so fare safely end-to-end encrypted. So far the process is used by US providers only. Abuse victim Alexander Hanff has been criticising chat control arguing „it will not stop children from being abused, it will simply push the abuse further underground, make it more and more difficult to detect and ultimately lead to more children being abused as the end result.“