European information sovereignty

First scoring edition · 10 November 2026 · EU-27 reference area

¹ Scores award points for observed EU origin or control. Unresolved cases earn zero in the headline and remain visible. If all unresolved cases qualified, the aggregate would move from 51.0 to 57.0. This is a classification range within the selected samples. Survey error, coverage and choice of weights add further uncertainty. This exploratory pillar uses observed shares. The other four pillars also use progress toward policy targets, so their scores have a different meaning.

Cinema / TV and music

Cinema counts tickets for EU films in 2023. Broadcast counts film minutes on 141 channels in 18 EU countries in 2023. Streaming counts EU film and series minutes on four services in nine EU countries in January–September 2024. Give each channel one vote because a common EU viewing-time denominator is unavailable. Film production origin is a proxy for content origin. European Commission, figure 25 · Observatory streaming sample.

Music uses the mean EU-artist representation in covered year-end national top tens in 2025. Each chart position gets one vote. Artist origin does not establish label ownership or the share of total listening. IFPI, Music in the EU 2026, pp. 10–11.

News

Reuters 2026 weekly brand reach. Take the top three entries per published market and medium, plus every tie for third. Retain anonymous outlet categories. Classify the controlling publisher or public institution, including the operator of licensed brands. Record divided or uncertain control separately. A public broadcaster qualifies on ownership irrespective of its editorial independence. Control labels are research judgements.

In each chart divide the sum of EU brands’ reach by the sum of all selected entries’ reach. People can name multiple brands. This yields the EU share of reported leading-brand uses. Average Belgium’s two published language charts equally (both return the same EU share). Weight the 20 markets by their population aged 18+ on 1 January 2025. They cover 97.7% of EU adults. Give online and offline equal weight. Smaller outlets, individual stories, time spent and seven EU states are outside this measure. Reuters report · Eurostat population inputs.

News score 77.8. The upper bound from the unresolved classifications is 93.1.

Social media for news and current affairs

Creator origin uses all 15 audience-nominated people or accounts in each of the six EU markets in Reuters’ comparable 2025 study. Score each market-list appearance equally. A recurring name counts in each audience where it appears. The proxy uses the creator’s identifiable professional or editorial base, inferred from public role and market profiles. Mobile or unclear cases remain unresolved. It does not trace where every reported story began. The survey prompted users of six named networks and included politicians and personalities. Global follower counts are unsuitable as weights for EU viewing. Reuters sampling method.

Of 90 list entries, 70 were coded EU, 6 non-EU and 14 unresolved. EU minimum 77.8%, upper bound 93.3%. The individual records below preserve the named platforms for the same accounts. Of the 70 EU-affiliated entries, 69 list only non-EU main services. This is a matched account-level observation. Only Spotify appears among the listed services with an EU-based parent. The survey’s six prompted networks were all non-EU controlled, so a separate survey supplies the platform score.

Platform origin uses European Parliament Flash Eurobarometer FL014EP, Q5 (June 2025). EU respondents who used social or video platforms for current affairs selected the services most important to them. All 12 explicitly offered services had non-EU parent organisations. “Other” accounts for 4 of 292 percentage points of service selections. Exclude the 1% “don’t know” response. The result is 0 to 1.4% EU platform mentions, using rounded published percentages. The options gave EU services only a write-in category. This is a weak instrument for measuring small services. EU survey data and questionnaire.

The two halves are independent proxies. Their mean is 38.9 to 47.4. It does not give the percentage of actual views that were both EU-created and EU-distributed. Creator lists give equal weight to names, even though audiences differ. Entertainment and personal feeds remain outside this score.

Reading the result

The model gives equal influence to four different activities. A different mix of viewing, listening and reading would produce a different result. Omitting the experimental social-media measure gives a three-area mean of 55.1. News and social media overlap where people use social platforms for news. The two indicators measure different features of that experience. Comparable EU audience minutes by creator and platform remain the largest data gap.

Additional ownership and classification evidence: Wirtualna Polska voting agreement · Fanpage ownership disclosures · Nordic ownership · PPF / CME · Ringier · Axel Springer ownership · Alpha TV co-ownership · Ströer shareholders · Rossel / 20 Minutes · ANSA cooperative · Mentour Pilot biography.