KEHA/73792/2024 - 14 October 2024

1. Decision

This decision is a decision pursuant to Section 28d of the Act on Information Management in Public Administration stating that the KEHA Centre will introduce an automated decision-making procedure for employment statements.

The decision-making procedure referred to in the decision will be used to issue employment statements to individual customers at the start and end of a job search, provided that the processing conditions for the automated decision-making procedure are met.

2. Justifications

2.1 Introduction

Automated decision-making concerning unemployment security is based on the Administrative Procedure Act (434/2003), the Unemployment Security Act (1290/2002) and the Act on Amending the Unemployment Security Act (1300/2023). In addition, the Act on Information Management in Public Administration (906/2019, hereinafter referred to as the Information Management Act) lays down provisions on the official responsibility and requirements for the deployment of automated decision-making procedures. These provisions ensure that the requirements of good governance are met, that accountability is correctly allocated, that decision-making is transparent and that automated decision-making is adequately controlled.

Where automated decisions target natural persons, the procedure must also meet the requirements of data protection legislation.

Statements issued in the unemployment benefit procedure are of such a nature that they can be resolved automatically under the Administrative Procedure Act and the specific law on the subject. The decision-making procedure complies with the requirements laid down in the Administrative Procedure Act and the Information Management Act, the Unemployment Security Act and, in the case of natural persons, the data protection legislation. The automated decision-making procedure has also been thoroughly documented and tested, and the necessary responsible persons have been appointed. The KEHA Centre has implemented the automated decision-making procedure and the decisions it automatically produces in accordance with the regulations and requirements.

The procedure fulfils the obligations imposed on the KEHA Centre and the employment authorities, including the legal principles of administration, the initiation of proceedings, the investigation of the case and the hearing of the party concerned, the resolution of the matter and the justification and notification of the administrative decision. Other principles of due process and good administration are also respected.

2.2 Conditions laid down in the Administrative Procedure Act

The conditions for automated decision-making are laid down in Chapter 8b of the Administrative Procedure Act, i.e. the chapter on automated decision-making. The chapter contains provisions on the conditions for automated decision-making, the condition for legal remedy and the obligation for the authority to inform the party concerned of the automated decision-making in the matter. The regulation gives the authority the power to take decisions automatically under the conditions set out in the chapter.

According to Section 53e of the Administrative Procedure Act, a decision is automated if the decision concluding the processing of the matter is made by means of automated data processing without being reviewed and approved by a natural person. Pursuant to Section 53e of the Administrative Procedure Act, an authority may automatically resolve a matter that does not include matters requiring case-by-case consideration. The matter may also be resolved automatically if the official or other person handling the matter has reviewed the facts of the case that require case-by-case consideration. Automated decision-making must be based on processing rules drawn up by the authority on the basis of the applicable law and prior discretion.

According to the explanatory memorandum to Section 53e of the Administrative Procedure Act (Government Bill 145/2022, p. 98), matters to be resolved automatically are, first of all, matters whose resolution can be derived in a sufficiently unambiguous manner directly from the law and the facts of the case. Secondly, matters can be resolved automatically where the legislation on which decisions are based includes a great deal of room for interpretation, but the authority has a well-established practice of deciding such matters, whereby matters with certain characteristics (type cases) are always resolved in the same way.

At the KEHA Centre and in the employment services, all procedures are based on legislation, and automated decision-making is also based on rules and regulations. Legislation refers to regulation in a broad sense, i.e. from EU law to lower-level standards and case law. The KEHA Centre has considered in advance which of the issues covered by this deployment decision can be resolved automatically, resulting in a legal resolution. Some of the matters that are resolved automatically are those where the resolution can be derived unambiguously from the legislation and the facts of the matter in question. In some matters, the KEHA Centre and the employment authorities have a well-established decision-making practice where cases are always resolved in the same way.

Automated decision-making is based on predefined processing rules. According to the Government Bill (145/2022 p. 106), processing rules refers to rules drawn up in advance by a natural person to guide the processing and resolution of an administrative matter by automated data processing, such as which variables influence the course of the operational process, how and on what grounds any interim measures are taken in the processing of the matter, at which stage of the operational process a human intervention is required, and how the resolution of the matter is formed in the operational process. The processing rules also include the rules under which a matter is included in the automated operational process or removed from it for processing by a natural person.

The processing rules also include the selection criteria determining whether a matter is processed automatically or selected for partly human processing where an official reviews and resolves the aspects that require case-by-case consideration. The processing rules have been drafted taking into account the provisions according to which the KEHA Centre and the employment authority must examine the information and explanations received in a manner justified by the nature and scope of the matter , the uniform treatment of jobseekers and the need for oversight. If the jobseeker is given the opportunity to be heard before the matter is decided and makes a spontaneous statement, the statement will be referred to an official.

The selection criteria are harmonised and based on a continuous risk analysis carried out by the KEHA Centre. In addition, inconsistencies in reporting are identified in some selections. Selection criteria may include, for example, discrepancies between the information declared and other information available to the KEHA Centre and the employment authority.

In its decision-making, the KEHA Centre only uses rules-based logic, and not, for example, logic based on probabilistic reasoning (so-called learning AI techniques). An automated decision is therefore always based on rules defined by law or derived from the law, in accordance with the established decision-making practices of the KEHA Centre and the employment authority, which can be programmed into the KEHA Centre's information systems as processing rules.

Section 53f of the Administrative Procedure Act provides for a condition of legal remedy. It is an additional prerequisite for a matter concerning natural persons to be resolved automatically. According to the condition of legal remedy, a decision concerning a natural person can be made automatically only if the person subject to the decision can claim full redress by means of an appeal to the authority which took the decision or an authority belonging to the same data controller. However, this is not required if an automated resolution is made to approve a claim of one party that does not concern another party.

A natural person cannot lodge an appeal against an employment statement, but can ask the employment authority to reconsider the statement or appeal against an appealable decision made by Kela or an unemployment insurance fund (hereinafter, the unemployment benefit payer) on the basis of the employment statement.

According to Chapter 12, Section 1 of the Unemployment Security Act, a jobseeker may lodge an appeal against an appealable decision that they have requested to correct a decision concerning the payment of unemployment benefit. Appeals are heard by the Unemployment Security Appeals Board and, in case of dissatisfaction with the decision of the Unemployment Security Appeals Board, by the Insurance Court.

If automated decision-making has been applied in making an administrative decision, the decision must indicate that the matter has been resolved automatically, in accordance with Section 53g of the Administrative Procedure Act, as well as the place where the deployment decision referred to in Section 28d of the Information Management Act is available. If an administrative decision is not issued on the basis of other legislation, the information must be provided to the party concerned by other means at the latest when the processing of the matter ends.

An automated resolution made on the basis of this deployment decision shall include an indication that the matter has been decided automatically within the meaning of Section 53e of the Administrative Procedure Act. The decision shall specify the address where this deployment decision is available. The decision is available in the public information network, on the Job Market Finland website.

2.3 Prerequisites laid down in the Information Management Act

Chapter 6a of the Information Management Act contains provisions that promote the principles of good administration. The provisions concern, among other things, the documentation of the automated operational process, quality assurance, quality control and the deployment of the procedure. The chapter also contains provisions on how the official responsibility for automated decision-making is allocated. It also contains provisions on how to ensure that the information used in the decision-making process is up-to-date and accurate, and on the obligation to provide information on the automated decision-making process.

As required by Section 28a of the Information Management Act, the KEHA Centre has recorded in a permanent format the responsibilities and background of the automated decision-making procedure under this decision. In drawing up the processing rules, the KEHA Centre has ensured that automated decision-making is lawful and meets the requirements of both substantive and procedural law. In connection with the drafting and documenting the processing rules, it has been made clear how non-discrimination will be ensured in the automated decision-making procedure. The processing rules also set out how the matter is adequately investigated and how the decision is justified or why it can be left unjustified. The KEHA Centre has also taken into account in the processing rules that if a matter requires case-by-case consideration, the processing of that aspect of the matter must be transferred to an official. The KEHA Centre has ensured that, for at least five years after a case has been closed, it can be shown under which processing rules an individual case was automatically resolved and at which stages an official was involved.

Following the documentation of the processing rules, the KEHA Centre has ensured, in accordance with Sections 28b and c of the Information Management Act, that the automated decision-making procedure implemented is in accordance with the documentation and that its quality is monitored systematically. The KEHA Centre has also ensured, for example through linguistic support, that good language is used in the documents provided to jobseekers and other interested parties, even when the documents are produced automatically.

The KEHA centre also carries out tests to ensure that the automated decision-making procedure meets the requirements. The necessary quality assurance and quality control plans have been drawn up for testing and quality control. They describe how to monitor the quality of the automated decision-making procedure and the correctness of automated resolutions, and how to take corrective action if an error is detected in the automated decision-making procedure. Persons responsible for ensuring the quality of the automated decision-making procedure, quality control and handling error situations have been appointed.

This deployment decision will be published in the public information network on the Job Market Finland website at tyomarkkinatori.fi as required by Section 28e of the Information Management Act. On the same web page, the KEHA Centre also provides information on the automated resolution of matters in its field of activity, the criteria for the use of the automated decision-making procedure and other information relevant to the rights of the customer based on the documentation and deployment decision referred to in Section 28a(2) of the Information Management Act.

2.4 Data protection legislation

Automated decision-making concerning matters relating to natural persons must also meet the requirements of the General Data Protection Regulation (EU 2016/679).

The GDPR requires national legislation on automated individual decisions. The legal basis for this legislation is Chapter 8b of the Administrative Procedure Act and Chapter 6a of the Information Management Act. Legislation sets out the safeguards to be respected in automated decision-making. Unemployment benefit statements and automated decision-making situations for individual customers are based on the law and meet the requirements of Article 22 of the General Data Protection Regulation.

The right to obtain human intervention

According to the General Data Protection Regulation, the safeguards of automated decision-making include, among other things, that the data subject (natural person) has the right to require human intervention in the processing of information. This safeguard is implemented by the condition of legal remedy under Article 53f of the Administrative Procedure Act, which states that the data subject may lodge an appeal against a decision taken automatically (see in more detail Chapter 2.2 Conditions laid down in the Administrative Procedure Act). 

A data subject cannot lodge an appeal against an employment statement, but can ask the employment authority to review the statement or appeal against an appealable decision made by an unemployment benefit payer on the basis of the employment statement. It is only from the decision thus obtained that the data subject can request an appealable decision to the extent that his or her claim has not been accepted. An appeal may be lodged against this appealable decision.

Informing the data subject about the automated decision-making procedure

EU data protection regulation and national law contain provisions that require the data subject to be informed when a decision concerning him or her is taken or may be taken automatically.

Articles 12, 13 and 14 of the GDPR provide for the data controller's obligation to provide information and related procedures. The GDPR stipulates that the controller must inform the data subject about automated decision-making. The purpose of the information regulation in the GDPR is to provide the data subject with a full and clear picture of the processing of personal data. This includes the obligation to inform the data subject in advance, at the initial stage of the process, that a decision on the matter can be taken automatically.

If automated decision-making has been applied in making an administrative decision, the decision must indicate that the matter has been resolved automatically, in accordance with Section 53g of the Administrative Procedure Act, as well as the place where the deployment decision referred to in Section 28d of the Information Management Act is available. The provision is one of the safeguards referred to in the General Data Protection Regulation. In order to ensure the legal rights of the person concerned in the automatic decision-making, the person concerned is informed that the decision in question has been made automatically. The fulfilment of the obligation to provide information has also been discussed above in Chapter 2.2 Conditions laid down in the Administrative Procedure Act.

The KEHA Centre has taken into account the requirements of the General Data Protection Regulation and the Administrative Procedure Act in its automated decision-making procedures. The KEHA Centre will provide the jobseeker with appropriate information indicating that their case can be processed in the automated decision-making procedure. In addition, the KEHA centre informs the jobseeker that human intervention will take place if the data subject lodges an appeal to request a change in a matter that has been resolved automatically.

Advance information

If a jobseeker starts using the services in a matter that can be resolved automatically, the KEHA Centre will inform the jobseeker about the automated decision-making in advance, right at the start of the process. When a jobseeker starts filling in a declaration or application in the e-service of the KEHA Centre and the employment authority, i.e. the Job Market Finland, they will be informed that the matter can be resolved automatically. The Job Market Finland text also states that more information on automated decision-making can be found on the Job Market Finland website at tyomarkkinatori.fi.

Other safeguards

The Administrative Procedure Act and the Unemployment Security Act provide for the procedural provisions for issuing unemployment benefit statements, such as the investigation of the case, hearing of the party concerned, the content and justification of the decision, correction of errors, the notice of appeal, notification, request for an appealable decision and request for rectification. These provisions shall also apply to automated decision-making. Procedural provisions, e.g. on the right to be heard and to receive a justified decision, as well as rules on appeals and official liability, are further safeguards related to automatic decision-making.

As a joint controller, the KEHA Centre has also fulfilled other obligations under the General Data Protection Regulation when processing personal data.

Special provisions concerning children

The automated decision-making procedure for job-seeking is not specifically targeted at children, but applies to people in the same way, regardless of whether they are of full age or under age. Unemployment benefit statements are always based on the law, and the principle of equality must be respected when issuing statements and taking decisions.

2.5 Conditions laid down elsewhere in legislation

The KEHA Centre's automatic decision-making procedure is based on the general regulations of the Administrative Procedure Act and the Information Management Act. Section 53f(1) of the Administrative Procedure Act limits the automated decision-making procedure in decisions concerning individual customers to cases where a person can seek an appeal or where the decision approves a claim made by the person.

3. List of documents on which the deployment decision is based

List of documents on which the deployment decision is based

• Processing rules for automated decision-making
• Quality assurance documents
• Documents for quality control and handling of errors

4. Retention period

If the automated decision-making procedure is discontinued, the KEHA Centre will retain this deployment decision for at least five years afterwards. The period starts from the beginning of the year following the discontinuation (Section 28d(3) of the Information Management Act).

5. Appeals

No appeal may be lodged against this deployment decision (Administrative Judicial Procedure Act, Section 6(2)). This document has been approved in the electronic case management system of the agency in question. The matter was presented by Matias Vainio, Head of Unit, and approved by Jaakko Westerlund, Director of Department.