prEN 18229-3
(Main)Artificial Intelligence – AI system requirements – Part 3: Logging
- Abstract
This document provides terminology, concepts, requirements, and guidance for logging of AI systems.
It is primarily intended for organizations placing on the market or putting into service AI systems and is not specific to any particular sector.
- Status
- Not Published
- Publication Date
- 25-Apr-2027
- Technical Committee
- CEN/CLC/JTC 21 - Artificial Intelligence
- Drafting Committee
- CEN/CLC/JTC 21/WG 4 - Foundational and societal aspects
- Current Stage
- 4060 - Closure of enquiry - Enquiry
- Start Date
- 20-Aug-2026
- Due Date
- 07-May-2025
- Completion Date
- 20-Aug-2026
- Directive
- 2024/1689 - EU AI Act
prEN 18229-3 is a CEN draft European Standard for logging in AI systems. It sets terminology, concepts, requirements and guidance for recording AI system events, mainly for organizations that place AI systems on the market or put them into service.
What does prEN 18229-3 specify?
prEN 18229-3 specifies how AI systems should log events across their life cycle, and how that logging supports traceability, post-market monitoring, deployer monitoring and, for RBI systems, biometric record keeping. The document is not limited to one sector, so it can apply to a wide range of AI products and services.
The document is organized into:
- Scope
- Normative references
- Terms and definitions
- Abbreviations
- Record-keeping
- Documentation
- Annex ZA on the link to Regulation 2024/1689
Annex ZA is informative and connects the standard to the essential requirements of the AI Act.
What are the key requirements of prEN 18229-3?
prEN 18229-3 centers on automatic logging, traceability, and evidence for later review. The provider has to decide what to log, how much detail to retain, and how the logs can support regulatory and operational needs.
Automatic logging and traceability
Clause 5.2 requires the AI system to have the technical ability to automatically record events throughout its life cycle. That matters because logging is treated as a system capability, not an after-the-fact manual activity.
The provider has to choose a traceability level that fits the intended purpose and the characteristics of hazardous situations. Table 1 then links those characteristics to extra logging needs, such as retaining logs until harm can be identified and attributed, logging each inference with inputs and outputs, or enabling aggregated queries. In practice, this means the provider must design logs so they can actually support investigation of fast-moving, hidden, or rights-related harm.
The provider also has to make sure the logs can identify both the affected version and the version that was conformity assessed. Clause 5.2.3 allows this either through the unique identifiers in FprEN 18286 or through an alternative mechanism that links logs to both identifiers. That matters for corrective actions, recalls, and version tracking through the supply chain.
How events are chosen for logging
Clause 5.3 says the provider must identify relevant events from risk scenarios, risk control measures, residual risk, system functions, interactions with other systems, human oversight functions, and the context of use. This ties logging to actual system behavior and use conditions rather than to a generic list.
Clause 5.3.2 requires continual reassessment after market placement or service. The provider updates the list of logged events using pre-market and post-market information and post-market monitoring. In practice, logging is not fixed once at launch - it must evolve with experience.
How logging is implemented
Clause 5.4 requires the logging information to align with FprEN ISO/IEC 24970. The provider must define storage and retention requirements, keep logs for at least six months or longer if required by the traceability conditions, and protect logs across their life cycle. This is the part that turns logging from a concept into an operational record system.
Clause 5.4.3 also requires the provider to minimize personal data in logs where possible without weakening traceability or the ability to identify hazardous situations or substantial modification. The document points to de-identification and metadata analysis as ways to reduce unnecessary personal data. That is important for privacy and data protection decisions during implementation.
Which events must be considered
Clause 5.5 groups events by purpose. The standard requires logging of specific events that support analysis of hazardous situations, identify possible substantial modifications, feed post-market monitoring, and support the deployer in monitoring system operation.
The document explicitly requires logging of events such as:
- potential attacks
- operation outside predefined limits
- adversarial attacks
- user requests
- requests from affected persons
It also requires systematic evaluation of other event triggers from FprEN ISO/IEC 24970 to decide whether they help identify hazardous situations, ineffective risk controls, or substantial modification. This means the provider has to justify exclusions, not just select a few convenient events.
For post-market monitoring, the provider must log events such as outlier inputs, safety events, performance events, unwanted bias detection, out of domain inputs, model drift and human oversight events. For deployer support, the provider must also evaluate errors, safety events, data quality issues and invalid data.
Additional rules for RBI systems
Clause 5.6 adds extra record-keeping for high-risk RBI systems. The logs must cover each operational cycle or use and preserve operational use, reference database information, input data and the natural persons involved in review.
For one-to-many biometric identification, the system must preserve the complete reference database as it existed at the time of identification, or a documented metadata-based alternative if full storage is technically impractical. For many-to-one biometric identification, the log must capture candidate match records, database metadata and protected biometric reference data in a way that prevents later alteration or deletion.
The clause also requires logging the original input biometric data for each potential match and recording manual review decisions by authorized individuals. In practice, this gives a reconstructable trail for biometric decisions, matching evidence, and human review.
Documentation and instructions for use
Clause 6 requires technical documentation for the logging capabilities and instructions for use that tell the deployer what is logged, how to access the logs, and how long logs are retained. Where provider access is technically difficult, the instructions must describe alternative arrangements such as user-retained logs or incident-triggered provision. That helps align the system design with the deployer’s obligations.
What terms does prEN 18229-3 define?
prEN 18229-3 defines key roles and concepts used throughout the logging requirements.
- provider - the natural or legal person that develops the AI system, or has it developed, and places it on the market or puts it into service under its own name or trademark.
- deployer - the person or body using the AI system under its authority, except for personal non-professional use.
- AI system - a machine-based system that can operate with varying autonomy and infer outputs from inputs to influence physical or virtual environments.
- intended purpose - the use intended by the provider, including context and conditions of use, as stated in the instructions for use, promotional material and technical documentation.
- substantial modification - a post-market change that was not foreseen or planned in the original conformity assessment and that affects regulatory compliance or intended purpose.
- event - a change of state or circumstances over time, including expected changes that do not occur.
- traceability - the ability to reconstruct the AI system and its history, including enough information from logs to verify conformity and intended purpose boundaries.
- RBI - remote biometric identification, the biometric use case that gets extra logging rules in Clause 5.6.
Who uses prEN 18229-3?
prEN 18229-3 is for organizations that place AI systems on the market or put them into service, especially providers that need to design logging into the system. It is also relevant to deployers because the instructions for use tell them how to access logs and how to use them for monitoring.
The standard is useful for AI product developers, compliance teams, quality managers, risk managers, cybersecurity teams, and documentation writers. It also matters to organizations operating biometric identification systems, because Clause 5.6 adds specific record-keeping rules for RBI systems.
Which standards are used with prEN 18229-3?
prEN 18229-3 refers to three main standards in its normative references:
- FprEN 18286 - supports quality management for EU AI Act regulatory purposes and provides the unique identifiers used for version traceability.
- FprEN ISO/IEC 24970 - provides the logging framework used for the content, storage, protection and event categories referenced throughout the document.
- prEN 18228 - provides the AI risk management basis used to identify risk scenarios, hazardous situations, risk control measures and post-market review.
The document also cites ISO/IEC 9868:2025 in the RBI guidance for biometric system workflows and manual review context.
What does the prEN 18229-3 document contain?
prEN 18229-3 contains two summary tables that are central to implementation. Table 1 links hazardous situation characteristics to additional traceability requirements, and Table 2 groups event triggers by the purpose of logging, such as risk, substantial modification, post-market monitoring and deployer monitoring.
The RBI clause contains worked examples and structured logging elements for database metadata, match records, archived input data and manual review actions. Those examples show how logs can be structured for biometric identification and review workflows without reproducing the underlying data formats.
Annex ZA contains the formal correspondence table between the standard’s clauses and the essential requirements of Regulation 2024/1689. That is the part readers use when they need to see which clauses support record-keeping and logging obligations under the AI Act.
Get Certified
Connect with accredited certification bodies for this standard

BSI Group
BSI (British Standards Institution) is the business standards company that helps organizations make excellence a habit.

Bureau Veritas
Bureau Veritas is a world leader in laboratory testing, inspection and certification services.

DNV
DNV is an independent assurance and risk management provider.
Sponsored listings
Frequently Asked Questions
prEN 18229-3 is a draft published by the European Committee for Standardization (CEN). Its full title is "Artificial Intelligence – AI system requirements – Part 3: Logging". This standard covers: This document provides terminology, concepts, requirements, and guidance for logging of AI systems. It is primarily intended for organizations placing on the market or putting into service AI systems and is not specific to any particular sector.
This document provides terminology, concepts, requirements, and guidance for logging of AI systems. It is primarily intended for organizations placing on the market or putting into service AI systems and is not specific to any particular sector.
prEN 18229-3 is classified under the following ICS (International Classification for Standards) categories: 35.030 - IT Security; 35.240.01 - Application of information technology in general. The ICS classification helps identify the subject area and facilitates finding related standards.
prEN 18229-3 is associated with the following European legislation: EU Directives/Regulations: 2024/1689; Standardization Mandates: M/593, M/593, M/613, M/XXX AI, M/XXX AI. When a standard is cited in the Official Journal of the European Union, products manufactured in conformity with it benefit from a presumption of conformity with the essential requirements of the corresponding EU directive or regulation.
prEN 18229-3 is available in PDF format for immediate download after purchase. The document can be added to your cart and obtained through the secure checkout process. Digital delivery ensures instant access to the complete standard document.
Standards Content (Sample)
SLOVENSKI STANDARD
01-julij-2026
Okvir zaupanja v umetno inteligenco - 1. del: Logiranje
AI trustworthiness framework - Part 1: Logging
Rahmenwerk für Vertrauenswürdigkeit von Künstlicher Intelligenz
Cadre de confiance pour l’IA
Ta slovenski standard je istoveten z: prEN 18229-1
ICS:
35.030 Informacijska varnost IT Security
35.240.01 Uporabniške rešitve Application of information
informacijske tehnike in technology in general
tehnologije na splošno
2003-01.Slovenski inštitut za standardizacijo. Razmnoževanje celote ali delov tega standarda ni dovoljeno.
EUROPEAN STANDARD DRAFT
prEN 18229-1
NORME EUROPÉENNE
EUROPÄISCHE NORM
May 2026
ICS 35.240.01
English version
AI trustworthiness framework - Part 1: Logging
Cadre de confiance pour l'IA Rahmenwerk für Vertrauenswürdigkeit von
Künstlicher Intelligenz
This draft European Standard is submitted to CEN members for enquiry. It has been drawn up by the Technical Committee
CEN/CLC/JTC 21.
If this draft becomes a European Standard, CEN and CENELEC members are bound to comply with the CEN/CENELEC Internal
Regulations which stipulate the conditions for giving this European Standard the status of a national standard without any
alteration.
This draft European Standard was established by CEN and CENELEC in three official versions (English, French, German). A
version in any other language made by translation under the responsibility of a CEN and CENELEC member into its own language
and notified to the CEN-CENELEC Management Centre has the same status as the official versions.
CEN and CENELEC members are the national standards bodies and national electrotechnical committees of Austria, Belgium,
Bulgaria, Croatia, Cyprus, Czech Republic, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Iceland, Ireland, Italy,
Latvia, Lithuania, Luxembourg, Malta, Netherlands, Norway, Poland, Portugal, Republic of North Macedonia, Romania, Serbia,
Slovakia, Slovenia, Spain, Sweden, Switzerland, Türkiye and United Kingdom.
Recipients of this draft are invited to submit, with their comments, notification of any relevant patent rights of which they are
aware and to provide supporting documentation.Recipients of this draft are invited to submit, with their comments, notification
of any relevant patent rights of which they are aware and to provide supporting documentation.
Warning : This document is not a European Standard. It is distributed for review and comments. It is subject to change without
notice and shall not be referred to as a European Standard.
CEN-CENELEC Management Centre:
Rue de la Science 23, B-1040 Brussels
© 2026 CEN/CENELEC All rights of exploitation in any form and by any means Ref. No. prEN 18229-1:2026 E
reserved worldwide for CEN national Members and for
CENELEC Members.
prEN 18226-1:2026 (E)
9 Contents Page
10 European foreword . 3
11 Introduction . 4
12 1 Scope . 5
13 2 Normative references . 5
14 3 Terms and definitions . 5
15 3.1 Terms related to the EU AI Act . 5
16 3.2 Other terms . 7
17 3.3 Terms related to risk . 9
18 4 Abbreviations . 10
19 5 Record-keeping . 10
20 5.1 Concept of event . 10
21 5.2 General principles for logging capabilities. 10
22 5.3 Requirements for identifying events to log . 12
23 5.3.1 General. 12
24 5.3.2 Continual reassessment of events to log . 12
25 5.4 Requirements for logging implementation . 13
26 5.4.1 General. 13
27 5.4.2 Storage and retention requirements . 13
28 5.4.3 Log protection, life cycle coverage and privacy . 13
29 5.5 Events to log . 13
30 5.5.1 General. 13
31 5.5.2 Events supporting analysis of hazardous situations that can be relevant to AI systems
32 presenting a risk . 14
33 5.5.3 Other events supporting analysis of hazardous situations that can present risks to health,
34 safety or fundamental rights or that can be relevant to AI systems presenting a risk . 15
35 5.5.4 Events indicating potential substantial modifications. 15
36 5.5.5 Events for post-market monitoring . 15
37 5.5.6 Events for supporting deployer in monitoring system operation . 16
38 5.6 Record keeping for RBI systems . 16
39 5.6.1 General. 16
40 5.6.2 Period of operational use . 17
41 5.6.3 Reference database . 17
42 5.6.4 Input data. 18
43 5.6.5 Natural persons involved in review . 18
44 6 Documentation . 19
45 6.1 Technical documentation . 19
46 6.2 Instructions for use . 19
47 Annex ZA (informative) Relationship between this European Standard and the essential
48 requirements of Regulation 2024/1689 aimed to be covered . 20
49 Bibliography . 21
prEN 18226-1:2026 (E)
50 European foreword
51 This document (prEN 18229-1:2026) has been prepared by the Joint Technical Committee CEN-
52 CENELEC/JTC 21 “Artificial Intelligence”, the secretariat of which is held by DS.
53 This document is currently submitted to the CEN Enquiry.
54 This document has been prepared under a standardization request addressed to CEN-CENELEC by the
55 European Commission. The Standing Committee of the EFTA States subsequently approves these
56 requests for its Member States.
57 For the relationship with EU Legislation, see informative Annex ZA, which is an integral part of this
58 document.
prEN 18226-1:2026 (E)
59 Introduction
60 5.1 introduces the concept of an event, and 5.2 provides general requirements and links these general
61 requirements to levels of traceability. 5.3 provides general requirements for events to log and 5.4 provides
62 requirements on the implementation of logging.
63 5.5 links the specific events to the objectives of logging, with requirements for how to determine exactly
64 which events are logged.
65 5.6 provides additional requirements for RBI systems.
prEN 18226-1:2026 (E)
66 1 Scope
67 This document provides terminology, concepts, requirements, and guidance for logging of AI systems.
68 It is primarily intended for organizations placing on the market or putting into service AI systems and is
69 not specific to any particular sector.
70 2 Normative references
71 The following documents are referred to in the text in such a way that some or all of their content
72 constitutes requirements of this document. For dated references, only the edition cited applies. For
73 undated references, the latest edition of the referenced document (including any amendments) applies.
74 FprEN 18286:—, Artificial intelligence — Quality management system for EU AI Act regulatory purposes
75 FprEN ISO/IEC 24970:—, Artificial intelligence — AI system logging
76 prEN 18228:—, AI Risk Management
77 3 Terms and definitions
78 For the purposes of this document, the following terms and definitions apply.
79 ISO and IEC maintain terminology databases for use in standardization at the following addresses:
80 — ISO Online browsing platform: available at https://www.iso.org/obp/
81 — IEC Electropedia: available at https://www.electropedia.org/
82 3.1 Terms related to the EU AI Act
83 3.1.1
84 provider
85 natural or legal person, public authority, agency or other body that develops an AI system (3.1.5) or a
86 general-purpose AI model or that has an AI system (3.1.5) or a general-purpose AI model developed and
87 places it on the market or puts the AI system (3.1.5) into service under its own name or trademark,
88 whether for payment or free of charge
89 Note 1 to entry: A distributor, importer, deployer (3.1.2) or other third party can be considered a provider (3.1.1)
90 of an AI system (3.1.5) in certain circumstances.
91 [SOURCE: EU AI Act 2024/1689, (Article 3(3)), modified – removed “a”]
92 3.1.2
93 deployer
94 natural or legal person, public authority, agency or other body using an AI system (3.1.5) under its
95 authority except where the AI system (3.1.5) is used in the course of a personal non-professional activity
96 [SOURCE: EU AI Act 2024/1689, (Article 3(4)), modified – removed “a”]
97 3.1.3
98 life cycle
99 evolution of a system, product, service, project or other human-made entity, from inception through
100 retirement
101 [SOURCE: ISO/IEC/IEEE 15288:2023, 3.21, modified – replaced “conception” with “inception”]
prEN 18226-1:2026 (E)
102 3.1.4
103 instructions for use
104 information provided by the provider (3.1.1) to inform the deployer (3.1.2) of, in particular, an AI system
105 (3.1.5) ’s intended purpose (3.1.6) and proper use
106 [SOURCE: EU AI Act 2024/1689, (Article 3(15)), modified – removed “means the”]
107 3.1.5
108 AI system
109 machine-based system that is designed to operate with varying levels of autonomy and that can exhibit
110 adaptiveness after deployment and that, for explicit or implicit objectives, infers, from the input it
111 receives, how to generate outputs such as predictions, content, recommendations, or decisions that can
112 influence physical or virtual environments
113 Note 1 to entry: The verb “can” represents a possibility, not all AI systems that fit the above definition have this
114 ability to adapt after deployment.
115 [SOURCE: EU AI Act (Article 3(1)), modified - removed “a” , replaced “may” with “can” based on the use
116 of verbs in standards. Added Note 1 to entry.]
117 3.1.6
118 intended purpose
119 use for which an AI system (3.1.5) is intended by the provider (3.1.1), including the specific context and
120 conditions of use, as specified in the information supplied by the provider (3.1.1) in the instructions for
121 use (3.1.4) promotional or sales materials and statements, as well as in the technical documentation
122 [SOURCE: EU AI Act 2024/1689, (Article 3(12)), modified – removed “means the”]
123 3.1.7
124 performance
125 ability of an AI system (3.1.5) to achieve its intended purpose (3.1.6)
126 [SOURCE: EU AI Act 2024/1689, (Article 3(18)), modified – removed “the”]
127 3.1.8
128 substantial modification
129 change to an AI system (3.1.5) after its placing on the market or putting into service which is not foreseen
130 or planned in the initial conformity assessment carried out by the provider (3.1.1) and as a result of which
131 the compliance of the AI system (3.1.5) with the applicable regulatory requirements is affected or results
132 in a modification to the intended purpose (3.1.6) for which the AI system (3.1.5) has been assessed
133 Note 1 to entry: Rephrased from the AI Act Article (3)(23) to reference applicable regulatory requirements instead
134 of a specific reference to the AI Act Chapter III, Section 2.
135 [SOURCE: EU AI Act 2024/1689, (Article 3(23)), modified – as described in Note 1]
136 3.1.9
137 input data
138 data provided to or directly acquired by an AI system (3.1.5) on the basis of which the system produces
139 an output
140 [SOURCE: EU AI Act 2024/1689, (Article 3(33)), modified – removed “means”]
prEN 18226-1:2026 (E)
141 3.1.10
142 personal data
143 information relating to an identified or identifiable natural person (also data subject)
144 3.1.11
145 biometric data
146 personal data (3.1.10) resulting from specific technical processing relating to the physical, physiological
147 or behavioural characteristics of a natural person, such as facial images or dactyloscopic data
148 [SOURCE: EU AI Act 2024/1689, (Article 3(34)), modified – removed “means”]
149 3.1.12
150 AI system presenting a risk
151 product presenting a risk
152 product having the potential to affect adversely health and safety of persons in general, health and safety
153 in the workplace, protection of consumers, the environment, public security and other public interests,
154 protected by the applicable Union harmonization legislation, to a degree which goes beyond that
155 considered reasonable and acceptable in relation to its intended purpose (3.1.6) or under the normal or
156 reasonably foreseeable conditions of use of the product concerned, including the duration of use and,
157 where applicable, its putting into service, installation and maintenance requirements
158 Note 1 to entry: For the purposes of this document, AI systems presenting a risk shall only be in so far as they
159 present risks to the health or safety, or to fundamental rights, of persons.
160 Note 2 to entry: These risks are not necessarily identified by the provider (3.1.1) before placing the AI system (3.1.5)
161 on the market or into service.
162 [SOURCE: EU Regulation 2019/1020, (Article 3(19)), modified – removed “means a” and added Note 1]
163 3.2 Other terms
164 3.2.1
165 biometric probe
166 biometric sample or biometric feature set input to an algorithm for comparison to a biometric
167 reference(s)
168 Note 1 to entry: In some comparisons a biometric reference can potentially be used as the subject of the comparison
169 with other biometric references or incoming biometric samples used as the objects of the comparisons.
170 Note 2 to entry: Typically, in a comparison process, incoming biometric samples serve as the subject of comparisons
171 against objects stored as biometric references in a database.
172 [SOURCE: ISO/IEC 2382-37:2022, 37.03.14]
173 3.2.2
174 biometric identification
175 automated recognition of physical, physiological, behavioural, or psychological human features for the
176 purpose of establishing the identity of a natural person by comparing biometric data (3.1.11) of that
177 individual to biometric data (3.1.11) of individuals stored in a database
178 [SOURCE: EU AI Act 2024/1689, (Article 3(35)), modified – removed “means the”]
prEN 18226-1:2026 (E)
179 3.2.3
180 remote biometric identification system
181 AI system (3.1.5) for the purpose of identifying natural persons, without their active involvement,
182 typically at a distance through the comparison of a person’s biometric data (3.1.11) with the biometric
183 data (3.1.11) contained in a reference database
184 [SOURCE: EU AI Act 2024/1689, (Article 3(41)), modified – removed “means an”]
185 3.2.4
186 one-to-many biometric identification
187 checking whether a specific known individual (probe subject) appears in a reference database
188 EXAMPLE Law enforcement checks a suspect’s photograph against a database of known criminals to verify
189 presence or absence in the database.
190 3.2.5
191 many-to-one biometric identification
192 identifying which individuals in a stream of probe subjects match entries in a reference database
193 EXAMPLE Post event (3.2.8) investigation where the AI system (3.1.5) compares faces of people passing by
194 against a watchlist of wanted individuals.
195 3.2.6
196 operational use
197 continuous period during which the system is active and performing biometric identification (3.2.2)
198 3.2.7
199 event
200 change of states of any type or a change of a particular set of circumstances, within a time continuum
201 Note 1 to entry: An event can have a single occurrence or multiple occurrences or a set of related occurrences. When
202 discussing multiple events, they can form a specific sequence or causal chain. It is important to note that an event
203 can have one or more causes and can lead to one or more consequences.
204 Note 2 to entry: This also includes an expected change of circumstances or of states which does not occur.
205 Note 3 to entry: Events are detected throughout the AI system life cycle (3.1.3): from inception based on the intended
206 purpose (3.1.6), through product development and field tests, to conformity assessment, to post-market surveillance
207 and user feedback loops.
208 Note 4 to entry: An event can be part of a set of events useful for different analyses such as co-occurrence analysis,
209 causal chain analysis, sequence analysis, or other time series analysis.
210 3.2.8
211 automatic
212 automation
213 automated
214 pertaining to a process or system that, under specified conditions, functions without human intervention
215 [SOURCE: ISO/IEC 2382:2015, 2121282]
prEN 18226-1:2026 (E)
216 3.2.9
217 integrity
218 property of accuracy and completeness
219 [SOURCE: ISO/IEC 27000:2018, 3.36]
220 3.2.10
221 traceability
222 ability to trace the AI system (3.1.5) and its history
223 Note 1 to entry: Traceability includes information on how AI systems have been specified, developed, verified,
224 validated, operated, monitored and retired.
225 Note 2 to entry: in the context of Article 12 of the EU AI Act [1] traceability means the ability to reconstruct, from
226 the logs automatically generated by an AI system, a sufficient account of how the system was functioning to allow a
227 competent authority or provider (3.1.1) to verify that the AI system was operating in conformity with the version
228 placed on the market and within the boundaries of its intended purpose (3.1.6).
229 [SOURCE: FprEN 18286:—, 3.3.2, modified – added Note 2]
230 3.2.11
231 de-identification process
232 process of removing the association between a set of identifying attributes and the data principal
233 Note 1 to entry: De-identification includes the process of altering data, via modifying or removing data, so that
234 individuals or entities cannot be identified, directly or indirectly.
235 [SOURCE: FprEN ISO/IEC 24970:—, 3.11]
236 3.3 Terms related to risk
237 3.3.1
238 harm
239 injury or damage to the health of a person or groups of persons, or interference with fundamental rights
240 Note 1 to entry: For the purpose of this document, damage to property or the environment, and the disruption or
241 destruction of critical infrastructure, are considered harms when they can result in injury or damage to the health
242 of a natural person or groups of persons or interference with fundamental rights.
243 Note 2 to entry: Interference with fundamental rights can be tangible or intangible, physical, psychological, societal
244 or economic, irrespective of the rightsholder’s awareness, in accordance with EU law, including the EU Charter on
245 Fundamental Rights.
246 Note 3 to entry: Safety in product safety risk management standards is understood as the absence of unacceptable
247 risk. In the context of this document, safety refers to the protection from harm from the use of the AI system (3.2.1).
248 [SOURCE: prEN 18228:—, 3.6.3]
249 3.3.2
250 hazard
251 potential source of harm (3.3.1)
252 Note 1 to entry: Threats and vulnerabilities to the AI system (3.1.5) itself are hazards only if they are themselves a
253 source of harm (3.3.1).
254 [SOURCE: prEN 18228:—, 3.6.1]
prEN 18226-1:2026 (E)
255 3.3.3
256 hazardous situation
257 circumstance in which people, property or the environment is/are exposed to one or more hazard (3.3.2)
258 [SOURCE: prEN 18228:—, 3.6.2]
259 4 Abbreviations
260 RBI remote biometric identification
261 5 Record-keeping
262 5.1 Concept of event
263 Logs of an AI system record events throughout the life cycle of the system according to pre-defined criteria.
264 Events can include changes of states of the AI system and changes to the data (digital and analog) the AI
265 system is processing such as changes in:
266 a) changes of input sensor measurement values (e.g. temperature values in degree Celsius), changes of
267 digital input states (e.g. a device changing from “on” state to “off” state), changes in IoT data streams
268 or changes in inputs (e.g. alarm state, or any continuous measurement value converted into digital)
269 from connected systems or devices;
270 b) changes in performances of the AI system, changes in resource utilization (e.g. CPU, bandwidth) or
271 changes in accuracy of AI predictions;
272 c) changes in security status;
273 d) changes in data storage and management measurements, changes in data quality and integrity
274 metrics.
275 5.2 General principles for logging capabilities
276 5.2.1 The AI system shall be designed and implemented with the technical capability to automatically
277 record events throughout its life cycle.
278 The logging capabilities shall be designed to enable these purposes:
279 a) identifying situations that can result in the AI system presenting a risk or in a substantial
280 modification;
281 b) facilitating post-market monitoring;
282 c) monitoring the operation of the AI system by deployers;
283 d) for AI systems performing RBI, the specific requirements.
prEN 18226-1:2026 (E)
284 5.2.2 The provider shall:
285 a) determine the level (extent/degree) of traceability of the functioning of the AI system appropriate to
286 its intended purpose and evaluate the characteristics of identified hazardous situations in accordance
287 with Table 1 to determine which corresponding requirements apply:
288 Table 1 — Traceability levels
Characteristics of the hazardous situation Specific additional requirements
1) Is harm associated with any identified Log retention shall cover the period until the
hazardous situation irreversible? harm can be identified and attributed to th
...



