Germany's highest court rules that Google is not obligated to ensure websites are free from defamatory content before displaying links to them in search results
Context & Ripple Effects
Germany's highest court closes out a years-long line of cases treating search engines as pointers rather than speakers: Google may link to a defamatory page without vetting it first, and the remedy sits with whoever was defamed, not with the index. The ruling lands alongside a broader judicial pattern — the EU's top court later held that a delisting order need not apply beyond the bloc's borders (right-to-be-forgotten stays regional), and Australia's high court went further, holding that Google is not a publisher of what it merely surfaces (Australia's not-a-publisher reversal).
The reason this 2018 decision still matters is what came after it: by 2026, a German court found Google directly liable for defamatory statements inside its own AI Overviews (direct liability for AI Overviews) — the same jurisdiction flipping from conduit to author once Google writes the text instead of linking to someone else's.
First-order effects
- Defamation claimants in Germany must pursue removal after a link appears rather than forcing pre-publication screening, leaving Google's German search operations free to index without legal review.
- Google's cost of compliance drops: no obligation to vet billions of pages means the burden of policing reputation shifts to publishers and individuals filing takedown requests.
Second-order effects
- National courts weighing parallel claims get a template — Australia's high court reached the same not-a-publisher conclusion four years later, suggesting the German reasoning travels across jurisdictions.
- Publishers defamed in indexed content absorb monitoring costs themselves, strengthening demand for delisting mechanisms like the EU's right-to-be-forgotten process as the practical recourse.
Third-order effects
- The conduit immunity rests on Google only pointing at third-party text; the 2026 AI Overviews liability ruling shows the shield dissolving when the platform generates the wording itself, drawing a structural line between indexing and authorship.
- If that line hardens, generative-answer features carry publisher-grade exposure that plain links never did — giving platforms a legal incentive to keep attribution-heavy formats and giving regulators a clean hook for imposing editorial duties on AI outputs.
The trend: Courts are splitting platform liability along an indexing-versus-authorship line: linking stays protected while machine-generated answers pull search engines toward publisher status.