Transkripce
Česká verze

Transcription in Public Administration: Standards, Accessibility, and Legal Requirements

· 7 min read ·

Public administration bodies produce recordings from council meetings, court proceedings, administrative hearings, and parliamentary work. Some of these recordings are subject to legal obligations regarding archiving, accessibility, and publication. Transcription and captioning in this context are not merely technical aids — they are tools for meeting legislative requirements and a prerequisite for accessibility for people with hearing impairments. This overview maps which laws are relevant, which types of recordings the obligations apply to, and what a technical transcription solution must specifically fulfil. This is not legal advice — specific applicability should always be verified with the relevant authorities or legal counsel.

Recordings in Public Administration — Types and Scope

Public administration produces recordings in several distinct contexts, each with its own rules.

Municipal and regional council meetings are the most widespread type of public proceedings with recording obligations. Municipal legislation in many countries requires that minutes be prepared from council sessions, filed at the relevant office for public inspection, and published on the official notice board and website. The legal requirement is typically for written minutes — not necessarily for audio or video recordings. However, many municipalities voluntarily produce and publish recordings on their websites. This is precisely where accessibility requirements come into play: once a video from a council session is published on the website of a public authority, accessibility obligations under applicable legislation apply.

Court and administrative proceedings follow different rules. Protocols documenting the course of court hearings are a legal obligation under procedural codes. Courts produce protocols capturing the proceedings — historically processed by court stenographers, not automatic transcription. Audio recordings from court hearings are part of the case file and are subject to archival and disposal obligations. Administrative proceedings include the obligation to produce a protocol of oral hearings as part of the case file.

Parliamentary proceedings have their own tradition: legislative bodies produce stenographic records from plenary sessions. Stenographic records are a historically established standard combining specialised shorthand systems with subsequent expansion into full text. Automatic transcription is not a direct replacement — it differs in method and accuracy for legislative terminology. Committees and public hearings whose recordings are published online are subject to the same accessibility rules as council recordings.

Legal Requirements for Digital Content Accessibility

The key legal framework for the accessibility of recordings in the public sector is the EU Directive 2016/2102 on the accessibility of the websites and mobile applications of public sector bodies. EU member states have implemented this through national legislation.

This legislation obliges public authorities, budgetary organisations, state enterprises, and other entities defined in its text to ensure that their websites and mobile applications meet accessibility standards. The technical reference document is WCAG 2.1 (Web Content Accessibility Guidelines version 2.1) at the AA level — an international standard for web content accessibility.

What WCAG 2.1 Requires for Audio and Video

Criterion 1.2.2 (Level A) requires captions for all prerecorded audio included in visual recordings. Criterion 1.2.4 (Level AA) extends this obligation to live broadcasts — live captions. This is a significantly more demanding requirement because it requires real-time captioning, not post-processing. Criterion 1.2.5 (Level AA) requires audio description for prerecorded video with visual content.

For council meetings, this means: a video from a session published on a municipal website must have captions. If a municipality broadcasts a session live via the web, the stricter live captioning requirement applies.

Captions Versus Text Transcript

WCAG 2.1 distinguishes between captions and a text transcript. Captions are synchronised with the video: each utterance appears at the correct time, speakers are identified, and relevant sounds are described (laughter in the hall, applause). A text transcript is a full record of the spoken word as a document, without video synchronisation.

A transcript as a document fulfils part of the requirements for alternative access to content, but it does not replace captions where WCAG requires synchronisation. For a video recording from a council session, publishing a text transcript alongside the video is therefore insufficient — the video must have captions directly in the player or as an SRT or VTT file loaded via the HTML track element.

SRT (SubRip) and VTT (WebVTT) are standard caption file formats accepted by all major web browsers and video platforms. Both contain timecodes assigned to each caption sequence and are open — they do not require proprietary software.

Archiving and Long-Term Preservation of Recordings

Archival legislation defines the obligations of records creators — that is, public administration bodies — in preserving documents. Audio and audiovisual recordings from public proceedings are documents in the meaning of such legislation. The institution's retention schedule determines how long they must be preserved and their disposal classification: recordings from council meetings typically carry permanent retention or disposal after a specified period.

A text transcript as supplementary archival material has practical value: it enables full-text searching within the archive contents. Recordings from decades of council proceedings are more accessible to historians, journalists, or lawyers if a searchable text counterpart exists. The transcript does not replace the original recording — it is a reference aid.

For digital document formats intended for long-term archiving, the recommendation is to prefer open standards independent of proprietary software. For text documents, PDF/A is the standard archival format; plain TXT is the most resilient alternative. For structured transcripts with metadata (date, institution, speaker names, timecodes), XML or JSON offers a more logical structure for machine processing and indexing.

Data Sovereignty and Security Requirements

Public administration works with data of varying sensitivity, and this sensitivity directly affects which transcription approach is permissible.

Administrative proceedings contain personal data — names, addresses, financial circumstances, health information, or trade secrets. Classified information legislation defines categories of classified information (Restricted, Confidential, Secret, Top Secret). Any processing of such materials by a transcription system must correspond to the security classification of the data.

GDPR (Regulation EU 2016/679) fully applies to public administration bodies when processing personal data. Article 35 imposes the obligation to carry out a Data Protection Impact Assessment (DPIA) before processing that is likely to result in high risk. Sending a recording from administrative proceedings containing personal or sensitive data to the cloud server of an external provider is precisely the situation where a DPIA must take place and where the outcome may be negative.

Local Transcription as an Answer to Security Requirements

Local transcription — a model running on the institution's own infrastructure — eliminates data transfer to third parties. Data does not leave the internal network and the institution has full control over where and how it is processed. A transcription system offering local processing through models like Whisper processes recordings on the institution's own machine. Export in SRT and VTT formats then directly meets the technical requirements of WCAG 2.1 for captions — a single application fulfils both the accessibility requirement and the data sovereignty requirement.

Local transcription also has its limitations: the accuracy of a model running on one's own hardware may not match the results of cloud ensemble systems aggregating across multiple models. For innocuous public proceedings — a council meeting video without sensitive content — cloud transcription with higher accuracy may be a more acceptable solution. The decision depends on the classification of the specific content.

Decision Framework for the Public Sector

Practical division for public administration institutions:

Recordings from public council meetings without personal data of specific individuals: cloud transcription is generally acceptable, provided a data processing agreement (DPA) is concluded with the provider under Article 28 GDPR.

Administrative proceedings and meetings with personal data of individuals: consider local transcription or a cloud solution with a strict DPA, DPIA, and verified security of the provider's infrastructure.

Recordings classified as sensitive or restricted information: local transcription on certified infrastructure meeting applicable security requirements. Cloud transcription is not an option.

Conclusion

Transcription and captioning in public administration are not merely a matter of practicality — in some cases, they are a legal obligation. EU Directive 2016/2102 obliges public sector bodies to ensure accessibility of digital content including captioning of published video recordings. Archival legislation frames the preservation of these recordings. GDPR and classified information laws set conditions for the secure processing of sensitive data. Technical solutions for transcription and captioning can meet these requirements — the prerequisite is correct content classification and selection of the appropriate data processing approach.

Sources: