Finland - Helsinki - AI Register

Overview

The City of Helsinki AI Register is a public, searchable repository that documents algorithmic and AI systems used within the City’s services. The register presents a standardized, user-centred set of metadata for each system—covering purpose, scope, accountability, datasets, data processing, human oversight, risks and mitigations, and lifecycle status—and it provides a feedback channel for residents to comment on systems. The register and its companion White Paper ("Public AI Registers", Version 1.0, September 2020) were developed to operationalise transparency as a core principle for trustworthy public-sector AI and to enable civic participation, auditability and better internal governance of municipal AI deployments. Key deployments documented include a range of chatbots, recommendation engines, and logistics/management systems used by City divisions; each entry includes contacts and supplier information so that citizens and auditors can follow up. For City-hosted information about the register see City of Helsinki AI Register and the White Paper at Public AI Registers (White Paper, Sept 2020). (ai.hel.fi)

Definitions

For the purposes of the Register, the City adopts pragmatic definitions drawn from the White Paper and international guidance: "AI system" and "algorithmic system" denote software or processes that make predictions, classifications or recommendations by automated processing of data; "register entry" denotes the public documentation record for a specific system; "accountable owner" is the City division or identified individual responsible for the system’s outcomes; "datasets" include training, reference and operational data used by the system; and "human oversight" denotes the structured arrangements (roles, escalation paths, human-in-the-loop interventions) by which personnel review, correct or override algorithmic outputs. These definitions align with the register template proposed in the City’s White Paper. (ai.hel.fi)

Governance and Institutional Framework

The register is operated by the City Executive Office of Helsinki (Kanslia), which is the central administrative unit responsible for planning and coordinating cross-divisional policies. The City Executive Office is named as the operator and point of contact for the register (email: [email protected]) and it is responsible for maintaining the website, coordinating entries across divisions, and ensuring accessibility and privacy compliance for the register site. Individual City divisions retain operational accountability for the AI systems they deploy; the register requires identification of both the accountable business owner and the technical owner for each system. The White Paper recommends assigning a responsible manager, technical owner and, where needed, independent auditors to ensure documentation quality. The register therefore functions both as an external transparency tool and as an internal governance artefact that supports lifecycle management, procurement oversight and supplier governance. See the City Executive Office overview at City Executive Office. (hel.fi)

Key Focus Areas

The Register emphasises multiple thematic focus areas: purpose and societal impact (why a system is used and who it affects), accountability (named responsible owners and supplier relationships), datasets and data governance (what data is used and its provenance), human oversight and explainability (how outcomes are reviewed and how decisions can be understood), non-discrimination and fairness checks (tested biases and mitigations), risk assessment and mitigation measures (identified risks and remedial actions), lifecycle status (pilot, production, retired), and public participation (feedback channels and transparency measures). The White Paper offers a recommended documentation template that operationalises these focus areas and explains how registers can be used to surface potential harms and to invite public scrutiny. The register pages for deployed City systems include concrete examples of dataset descriptions, contact points and external suppliers. These focus areas are intended to support both democratic oversight and better internal risk management practices. (ai.hel.fi)

Implementation Framework

Implementation of the Register follows a practical template: (1) Each City division documents systems using a common entry template derived from the White Paper; (2) Entries are posted on the public site, with non-sensitive technical details and accessibility-conscious language; (3) Contact information is provided for accountable owners and suppliers; (4) The register includes lifecycle tagging and links to deeper technical or procurement documentation where permissible; (5) The City operates the site and coordinates updates via the City Executive Office and assigned digital teams. The White Paper encourages collaboration between public teams and external vendors to produce accurate entries and suggests that registers should be used as part of procurement and audit workflows. The register site includes accessibility and privacy notices that govern the public presentation of information and feedback mechanisms. (ai.hel.fi)

Monitoring and Evaluation

Monitoring is achieved through a combination of procedural and technical measures: division-level responsibilities for updating entries, internal audit or review mechanisms recommended by the White Paper, and public feedback channels that allow residents to flag concerns. The register can be used as an input to human-rights and data-protection impact assessments and to internal compliance checks. The City’s accessibility and privacy statements for the register explicitly note audit and ongoing inspection practices for the website itself. Over time, the register is expected to support evaluation of AI deployments across the City, enabling comparisons of risk profiles, supplier dependencies and mitigation measures. The White Paper suggests registers become part of routine documentation and review cycles. (ai.hel.fi)

Penalties, Liability, and Appeals

The AI Register is a transparency and governance policy of the City rather than a statutory sanctioning instrument. The register itself does not establish new criminal or administrative fines; instead, non-compliance with documentation expectations is addressed through internal administrative procedures and procurement/contractual remedies. Separate legal liability for data protection and fundamental-rights infringements remains governed by national and EU law, including the GDPR and administrative law: data protection authorities and courts retain powers to investigate and impose fines or corrective measures where applicable. The City’s public pages and privacy statements reference applicable data-protection oversight and encourage reporting through official channels. For the GDPR and EU legal framework see the official EU legislation resources. (ai.hel.fi)

Relationship to Other Instruments

The Register complements other City and national instruments: Helsinki’s own ethical principles and digital strategy documents guide responsible use of data and AI; the register functions as a transparency mechanism alongside procurement rules, privacy impact assessments and IT security policies. The White Paper explicitly positions public AI registers as operational complements to existing legal frameworks and to EU-level instruments such as forthcoming or proposed AI regulation. The register also interoperates with the City’s accessibility and privacy notices that set standards for public-facing documentation. See Helsinki’s ethics announcement and related City content for contextual policy alignment. (hel.fi)

International Alignment

The City’s approach is explicitly framed to align with international trustworthiness and transparency principles for AI. The White Paper cites EU-level guidance (e.g., high-level expert group documents) and positions the register as compatible with both the GDPR and international human-rights-based impact assessment approaches. Helsinki’s register was developed in collaboration with international partners and the vendor Saidot and has been referenced as an example for other municipalities. The White Paper was produced to encourage replication of the register model across jurisdictions and to harmonise transparency practices for public-sector AI. (ai.hel.fi)

Implementation Timeline

Sources and References