Author: AlgorithmWatch Switzerland
Source: algorithmwatch.ch – Approach AI Governance Holistically
Publication Date: 2025 (Statement)
Reading Time: approx. 5 minutes


Executive Summary

The Federal Council has announced a strategy for AI deployment in federal administration and adopted new guidelines on digital sovereignty – important signals in a still largely unregulated field. However, significant implementation gaps are evident: missing resource planning, vague specification of fundamental rights protection, and fragmented governance without overarching strategy. Without holistic coordination and financial resources, this initiative risks failing to realize its transformative potential.


Critical Key Questions

  1. Freedom & Fundamental Rights: How are citizen rights concretely protected when AI deployment is primarily pursued under efficiency objectives?

  2. Responsibility & Distribution: Who benefits from efficiency gains – and who bears the risks of algorithmic errors?

  3. Transparency & Implementation: Why do central implementation principles (sustainability, fundamental rights protection) remain vague and non-binding?

  4. Sovereignty & Dependency: How does the state effectively minimize its dependency on Big Tech corporations?

  5. Coherence & Strategy: Why is AI governance not linked to other regulatory levels (EU Convention, platform regulation)?


Scenario Analysis: Future Perspectives

Time HorizonExpected Development
Short-term (2026)Resource decision in spring 2026; if underfunded, protective measures will be delayed. Fragmented pilot projects without clear success criteria.
Medium-term (2028–2030)AI deployment shows efficiency gains but also reveals fundamental rights risks. Pressure for European harmonization grows (EU AI Act as reference).
Long-term (2035+)Either: integrated, transparent AI governance with genuine fundamental rights protection OR: fragmented solution with technological lock-in dependency on major providers.

Main Summary

Core Topic & Context

The Swiss Federal Council responds to the global AI boom with two initiatives: a strategy for AI in administration and new guidelines on digital sovereignty. However, implementation lags behind ambitions – resources unclear, principles imprecise, governance fragmented.

Key Facts & Figures

  • Federal Council plans implementation measures for AI strategy and strengthens internal coordination
  • Resource decision not until spring 2026 – critical delay
  • Focus thus far heavily on efficiency improvements, less on equal opportunities or sustainability
  • New digital sovereignty guidelines intended to minimize dependency on tech giants
  • ⚠️ Fundamental rights impact assessments are not mandatory in project management integration

Stakeholders & Affected Parties

  • Winners: Federal administration (efficiency gains), tech providers (increased demand)
  • Losers: Citizens without transparent control over algorithmic decisions; employees facing efficiency-driven cuts without redistribution plan
  • Observers: Data protection and fundamental rights organizations, European regulators, civil society

Opportunities & Risks

OpportunitiesRisks
Proactive action before EU regulatory pressureUnderfunding leads to half-heartedness
Digital sovereignty reduces vendor lock-inEfficiency gains without redistribution concept
Fundamental rights protection as governance principleVague implementation of sustainability & fundamental rights
Exemplary role for European standardsFragmentation without overarching strategy

Action Relevance

For Decision-Makers:

  • Immediately: Secure funding commitment for 2026, otherwise credibility is lost
  • Short-term: Establish binding concrete implementation requirements for fundamental rights protection and sustainability
  • Strategically: Link AI governance with platform regulation, EU standards (AI Act), and Council of Europe Convention
  • Monitoring: Establish mechanisms for regular review and public accountability

Quality Assurance & Fact-Checking

  • [x] Central statements verified (Federal Council communication May 2025)
  • [x] Resource decision timeline correct (spring 2026 confirmed)
  • [x] Unconfirmed details marked with ⚠️ (integration requirement for fundamental rights impact assessments)
  • [x] No identified one-sided bias – statement is critically constructive
  • [x] Structural gaps factually substantiated

Supplementary Research

  • Council of Europe AI Convention (2023): Binding standards for fundamental rights protection – Switzerland not yet ratified
  • EU AI Act (in force 2024/2025): De facto standard for European administrations – harmonization pressure on Switzerland increasing
  • Swiss Data Protection Act (revWG 2023): Foundation for algorithm transparency, but not yet operationalized for AI

References

Primary Source:
AlgorithmWatch Switzerland: "Approach AI Governance Holistically" – algorithmwatch.ch/de/ki-governance-ganzheitlich-angehen/

Supplementary Sources:

  1. Swiss Federal Council: Statement on AI Strategy and Digital Sovereignty (2025)
  2. Council of Europe: Convention on Artificial Intelligence (2023)
  3. European Commission: AI Act Implementation Guidance (2024)

Verification Status: ✓ Facts checked on 05.12.2025


This text was created with support from Claude 3.5 Sonnet.
Editorial responsibility: clarus.news | Fact-checking: 05.12.2025