On April 17, 2026, Federal Justice Minister Dr. Stefanie Hubig published the Referentenentwurf of the Act Against Digital Violence (Gesetz gegen digitale Gewalt), the long-promised replacement for Germany's Network Enforcement Act. Stakeholder consultation closed on May 22, and entry into force is expected later this year. The draft does something unusual for German platform regulation: instead of conscripting companies to police speech under threat of fines, it puts a judge in the loop.
NetzDG's flawed inheritance
The 2017 NetzDG required platforms with more than two million users to delete "manifestly unlawful" content within 24 hours and other illegal content within seven days, on pain of fines up to €50 million. The design was widely criticised — including by the UN Special Rapporteur on freedom of expression — for creating an incentive to over-remove: faced with a tight clock and a large fine, a rational platform deletes in case of doubt, sweeping up lawful speech with the unlawful. NetzDG was largely superseded by the EU Digital Services Act, and the new draft would formally repeal it.
A court-centric design
The draft's core move is to route enforcement through civil courts rather than corporate moderation teams:
- Judicial disclosure orders. A victim seeking a perpetrator's identity must obtain a court order; the platform then discloses IP data, and a second order compels the access provider to reveal subscriber details (name, date of birth, address, e-mail). The disclosed IP set is broader than before, covering the address used at the time of the infringement and at last login.
- Court-ordered account suspension. In severe personal-rights cases, a court may order temporary blocking of an account to prevent recurrence — explicitly a last resort, subject to proportionality and procedural safeguards.
- Evidence preservation. Once proceedings begin, providers must secure relevant data and content, transfer it if ordered, and irreversibly delete it after a final decision.
- Domestic service agent. Social networks without an EU establishment must appoint a German agent for legal proceedings; EU-based providers only upon a specific court order.
Three new criminal offenses accompany the civil framework: §184k StGB on non-consensual intimate imagery, including sexualized deepfakes and voyeuristic recordings; §201b StGB on defamatory fabricated content depicting real people, carrying up to two years' imprisonment; and §202e StGB on covert GPS or Bluetooth stalking.
The free-speech case for the new model
Start with the strongest argument for acting at all. Digital violence is not abstract. A 2024 TUM–HateAid study of roughly 1,100 politically active people found that about two-thirds of affected women had faced sexualized online attacks, and around 22% had considered abandoning political work entirely. When abuse drives people out of public life, the marketplace of ideas shrinks. A state that wants an open internet has a legitimate interest in keeping its participants from being silenced.
What is genuinely encouraging is the method. By replacing NetzDG's fine-backed deletion mandates with court orders, the draft removes the structural pressure to over-censor. A judge — not a moderation queue racing a 24-hour deadline — now decides whether speech is unlawful, whether anonymity should be pierced, and whether an account should be suspended. Suspension is framed as a last resort with proportionality review, and the draft even lets accused users participate anonymously or pseudonymously. This is closer to how a liberal legal order should handle contested speech: individualized, adversarial, and reviewable.
Where proportionality is at risk
The promise is real; the execution is where it can misfire. Three concerns stand out.
First, resourcing. The advocacy group HateAid warns that victims must front legal costs before any prospect of recovery, and the German Judges Association cautions that overstretched courts cannot absorb the new caseload — "three investigators doing the work of four." A remedy only the well-resourced can use, processed by courts too thin to move quickly, protects no one in practice.
Second, de-anonymization. Anonymity is itself a free-speech good: it shields whistleblowers, dissidents, and ordinary critics. Judicial supervision is the right safeguard, but the broadened IP-disclosure set and the likely volume of civil claimants mean the real test is how narrowly courts read "digital violence." Defined loosely, identity-unmasking becomes a tool to chill criticism rather than punish abuse.
Third, the DSA problem. The Digital Services Act deliberately centralised platform obligations at EU level under a country-of-origin rule, precisely to stop 27 national regimes from fragmenting the single market — the lesson of NetzDG. Berlin argues its measures escape that rule because they are individual judicial orders, not general regulatory duties. That is a clever characterization, but a contestable one; if it holds, other member states will copy it, and the harmonization the DSA bought will erode order by order.
The verdict
The Act Against Digital Violence is, on balance, a better instrument than the law it replaces. Moving from platform-as-censor to judge-as-arbiter is the proportionate, rights-respecting direction, and the new deepfake offenses fill genuine gaps. But a court-centric model is only as good as the courts behind it. If Germany funds the bench, defines digital violence tightly, and resists turning disclosure into a routine unmasking tool, this could be a model worth exporting. If it does not, it will have written elegant rights that few can afford to use — and reopened the single-market fragmentation the DSA was meant to close.