Versatile AI Risk Assessment
Office work and knowledge management

Machine translation

Public authorities and businesses automatically translate texts, websites and documents, often using their own glossary.

Explore the interactive model

Export handout

A concise PDF covering how the system works, its data flows and points to investigate, ready to share.

The explanations and measures describe the typical system design. Extensive personal changes add an appendix.

Understand the use case

What defines this use case?

The European Commission operates eTranslation, a city automatically translates its website, and the US border agency translates conversations and documents.

This page describes a typical system design. Actual data flows and permissions depend on your implementation.

System design
Generative assistance
Ability to act
Action permissions depend on integration
Fields of use
public administration, transport
Also known as
AI translation, neural machine translation

How it works, step by step

System design and data flows

  1. Input data

    A text, document or website arrives in the source language, optionally with a glossary.

    Correspondence, websites and statements from outsiders are also translated. Some systems also translate speech or text in images and scans.

  2. Processing

    The model translates sentences in context and follows glossary terms for specialist vocabulary and product names.

    The model processes the source text in its linguistic context and formulates a text in the target language. Terminology, relationships between sentences and tone are expressed anew. Glossaries or additional context can support specialist terminology and proper names, where the tool provides these features.

  3. Results

    The output is in the target language, optionally preserving the original layout.

    A generated result is newly worded or created. To assess it, the instruction and context on which it is based must remain identifiable.

  4. Use and impact

    The translation is published, sent or used to support decisions.

    Integration determines whether the result is simply displayed, adopted after approval or written directly into a business application. These variants differ in their data flows and permissions.

Explore the system design: Select a component to see its task, incoming and outgoing data, and role in the use case. The guided walkthrough explains each phase. Then use Yes, Uncertain and No to represent your specific system variant.

8 review first15 clarify8 baseline threats

InputProcessingResultImpact Externalaccess 04 Input 31 Third-partycontent 00 AI model 05 Providerservice 11 Documentcollection 00 Memory 01 Training andfeedback 11 OtherAI agents 00 Tools andinterfaces 00 Result 30 Business appor equipment 01 person 00 Affectedperson 00 AI serviceswithout approval 01 Externalaccess 04 Input 31 Third-partycontent 00 AI model 05 Providerservice 11 Documentcollection 00 Memory 01 Training andfeedback 11 OtherAI agents 00 Tools andinterfaces 00 Result 30 Business appor equipment 01 person 00 Affectedperson 00 AI serviceswithout approval 01
  • Present
  • Depends on your system
  • Absent
  • Review first
  • To clarify
Architecture as a list

How it differs from similar use cases

Data and information

Input data and information sources

Correspondence, websites and statements from outsiders are also translated. Letters, applications and border conversations identify people; product texts and websites usually do not.

Processing and operating environment

Translation models generally come from providers such as DeepL or the European Commission.

Storage and further use

Translation memories and glossaries reuse earlier translations in new texts. Some deployers adapt models using their own translations; most use them unchanged with a glossary.

Results and data sharing

Automatic website translation publishes outputs without further review. Cloud services send texts outside the organisation's environment.

Using a provider does not automatically determine where processing takes place. Storage location, subcontractors, deletion periods and use for training must be checked for the specific deployment.

Assess risks and determine measures

From understanding the system to analysing risks

  1. Define the system boundary

    Record data, participants, interfaces and permissions. Resolve uncertain characteristics in the model.

  2. Investigate threats

    Relate review candidates to actual data flows: what could happen, and who would be affected?

  3. Assess risk

    Assess impacts and likelihood in the deployment context. The number of review candidates is not a risk score.

  4. Provide evidence of measures

    Select appropriate controls, assign responsibilities and define effectiveness checks. Document residual risks and approvals.

Key threats in the application context

  • 3 particularly important
  • 5 typical
  • 15 depending on the system
  • 8 additional threats for every AI system

A total of 31 threats from the threat catalogue may be relevant to this use case.

Priority review points

  1. Factual Inconsistencies (Hallucinations)

    The model omits a negative or mistranslates a technical term while still sounding fluent. The meaning of a decision or instruction changes unnoticed.

  2. Sensitive Information Disclosure

    Staff copy contracts or personnel records into a translation service. Provider retention or reuse sends confidential content outside the organisation.

  3. Insecure Output Handling

    An automatically translated website goes live without review. Errors in deadlines, fees or responsibilities reach citizens or customers directly.

5 further review candidates in the typical design

Other relevant threats

  1. Targeted Poisoning / Label Poisoning

    Falsified training labels can deliberately shift model outputs.

    Listed because of this characteristic

    • Learns from examples with correct answers

  2. Backdoor ML Model

    Training data or an adopted base model can introduce hidden triggers for misbehaviour.

    Listed because of this characteristic

    • Uses another provider's model

  3. Model Denial of Service

    Complex or high-volume inputs can exhaust model computing resources.

    Listed because of this characteristic

    • Generates its own text

  4. Privacy Attacks

    Targeted queries can recover some personal or protected training data from the model.

    Listed because of this characteristic

    • Processes confidential data

  5. Exfiltration from ML Application

    Protected data can leak through tools, network connections, storage or insiders.

    Listed because of this characteristic

    • Processes confidential data

Up to 15 further threats may apply, depending on your system design. Nine threats need to be reviewed for every AI system. One of these is particularly important here and is therefore listed above.

System features and open questions

Each Yes adds further threats to your checklist, up to 15 in total. If you cannot yet answer a question, include its threats in your review.

How many threats a Yes adds
  • Added by this Yes
  • Also requires a Yes to another question

How your system is built

  1. Performs natural language tasks5 threats
  2. Learns from your own data1 threat, plus 1 only in combination with other characteristics
  3. Remembers content beyond a single conversation1 threat
  4. Changes data or triggers processes1 threat
  5. Uses operational feedback1 threat
  6. Processes text, images and speech together1 threat

How you operate the system

  1. Accepts queries directly through an interface4 threats
  2. Available to people outside the organisation4 threats
  3. Accessible over a network1 threat, plus 1 only in combination with other characteristics
  4. Incurs costs with each query1 threat
  5. Shares computing resources with other customers or applications1 threat
  6. Used alongside unapproved AI services1 threat

How your system is built

Does the system remember content from earlier conversations or tasks and reuse it later?

Translation memories and glossaries reuse earlier translations in new texts.

If yes, add

Can the system change things itself, such as storing data, sending messages, initiating transactions or granting permissions?

Automatic website translation publishes outputs without further review.

If yes, add

Does your organisation train or fine-tune the model itself using its own data?

Some deployers adapt models using their own translations; most use them unchanged with a glossary.

If yes, add

Does the system process several input types together, such as text with images or speech?

Some systems also translate speech or text in images and scans.

If yes, add

How you operate the system

These questions arise for almost every AI system.

Do staff also use AI services for this task that your organisation has not approved?

Without an approved service, staff copy texts into free online translators.

If yes, add

Controls and implementation

Measures for the priority review areas are expanded. Each linked measure explains how to implement it.

Review priorityFactual Inconsistencies (Hallucinations)3 measures
Review prioritySensitive Information Disclosure3 measures
Review priorityInsecure Output Handling3 measures
Targeted Poisoning / Label Poisoning3 measures
Backdoor ML Model2 measures
Model Denial of Service3 measures
Privacy Attacks3 measures
Exfiltration from ML Application3 measures

Controls for the remaining threats are available on their pages in the threat catalogue.

AI Act provisions

When AI Act provisions apply
  1. Art. 4: AI literacy

    For every AI system Applies since 2 February 2025.

  2. Annex III no. 7: Migration, asylum and border control

    Depending on use Review trigger from 2 December 2027.

Applies to every AI system

Art. 4: AI literacy

Applies if staff or appointed representatives operate or use AI systems.

Applies since 2 February 2025.

Obligations and details
Who bears the obligations
Provider, Deployer, Product manufacturer, Provider of a general-purpose AI model

Detailed obligations

May apply in particular uses

Annex III no. 7: Migration, asylum and border control

Applies if migration or border authorities use translations to examine asylum, visa or residence applications and assess the credibility of statements.

Applies from 2 December 2027.

Obligations and details
The general condition is
Applies if the system is used in examining asylum, visa or residence applications or at border controls.
Who bears the obligations
Provider, Deployer, Product manufacturer

Detailed obligations

Classify your system with the quick checker All obligations in the obligations compass

Cross-cutting threats

Review these threats for every AI system, whatever its purpose.

Why these threats always belong on the checklist

Atypical system features

View 13 further questions

If any of these features does apply to your system, add the listed threats.

Sources

Related use cases

Also in the field of Office work and knowledge management

Often used together with

Updated: 30 September 2026. Review checklist rules version 0.3.2, threat catalogue version 2026.10.24.2. How these pages are structured