Negotiators from the European Parliament and the Council of the EU reached an agreement, on Thursday 10 December, on the regulation for the removal of terrorist content online. This text will allow a competent national authority to request a platform or content host to remove violent material of a terrorist nature within one hour.
The crux of the negotiations in recent days has been the cross-border effect of withdrawal orders, i.e. when a national authority requests such a withdrawal from an actor located in another Member State.
Like the European Commission, the EU Council initially advocated a direct and immediate effect on the actor from another Member State. The European Parliament, on the other hand, saw this as an infringement of sovereignty and demanded an active right of scrutiny for the competent authority of the host country and a right of appeal for the company concerned.
According to the compromise reached, the authority of the Member State where the content host is located will have 72 hours to analyse and possibly challenge the withdrawal order, if it sees, for example, a clear violation of freedom of expression. But the content must be made inaccessible immediately. Depending on the decision that will be taken after the challenge, the content may either become available again or be permanently deleted.
The host or the author of the content will have 48 hours to contest the withdrawal order from the country of origin.
On the very controversial issue of the use of automated filters as a preventive measure, the interinstitutional compromise excludes mandatory measures and will leave the choice to the hosting or content providers.
“The decision as to the choice of specific measures shall remain with the hosting provider and the service provider”, says the compromise text.
Non-mandatory automated filters
These preventive measures may include one or more of the following elements: “appropriate technical and operational measures or capabilities, such as appropriate human or technical means to detect, identify and expedite the removal or deactivation of terrorist content”, the compromise states. If the action is deemed insufficient, the competent authority could again request the undertaking concerned to take appropriate action against the disputed content. Again, however, the choice of means would be left to the operator.
A sanction would be imposed after several unsuccessful requests for withdrawal. It will be up to the Member States to identify the nature of the penalty.
Greens/EFA negotiator Patrick Breyer (Germany) welcomed the fact that automated filters are not an imposed measure, which he sees as hidden censorship of the internet and a violation of fundamental rights.
He also welcomed more secure language on the protection of journalistic or academic work. The compromise now also clarifies that potentially controversial content aimed at fuelling public debate should not be considered as being terrorist content.
For the rest of the measures, the regulation also provides for accommodations for small businesses that would have logistical difficulties to remove or deactivate this content within the deadline.
Other European Parliament political groups and the Commission welcomed the agreement, which the EU Council is expected to endorse on 16 December. On 14 December, the EU Interior Ministers will already take note of this during their videoconference. EU leaders are also expected to welcome this outcome in the security conclusions they will adopt on Friday 11 December. The European Parliament could endorse the agreement in January 2021.
This issue had come back to the forefront in the wake of the attack on Samuel Paty in Conflans-Sainte-Honorine on 16 October.
Link to the draft compromise: https://bit.ly/3m35qZe (Original version in French by Solenn Paulic)