Contributor Consent

You are invited to contribute to BUDOVA ("Building Ukrainian Domain-Specific, Open Voice & Text Archives"), an open Ukrainian construction-domain language corpus for AI research run by the KNUCA AI Lab. This consent explains what data we process when you contribute, why, how your contributions will be published and what your rights are, under the Law of Ukraine "On Personal Data Protection" and, where it applies, the GDPR. The Privacy Policy gives the full detail.

Participation is entirely voluntary: you may decline, or stop at any time, with no consequences. Recording your voice is a further, separate choice. This is version 1.1 of the document (v1.1-2026-10-18). Your in-app consent is recorded with the full name you type as an e-signature, the time, this version, your IP address and browser, and whether you gave it at the consent gate or in the recorder.

This document is published in Ukrainian and in English with the same content. If the two versions differ, the Ukrainian version prevails.

The controller responsible for processing your data is:

Kyiv National University of Construction and Architecture (KNUCA), AI Laboratory — project BUDOVA
31 Povitryanykh Syl Ave., Kyiv, 03037, Ukraine
Data protection contact: dolhopolov_sy@knuba.edu.ua

• Account — email address and name; optionally a public display name and profile links.
• Contributions — your annotations of construction texts (with the time spent), texts and terms you submit, and texts you add for reading aloud.
• Voice recordings — only if you choose to record: your voice; what you read (and any wording or comment of your own); your speaker region (where your way of speaking comes from) and, if you give it, your town or village; technical measures of the audio; and a speaker code (SPK-NNN).
• Speaker profile — optional: age band, gender, experience in construction, home language and profession. Only the research team sees these answers; they are never published.
• Technical and security data — IP address, browser, sign-in and activity times, and audit-log entries of your consents, exports and withdrawals.
• Usage events — product analytics (PostHog, EU cloud) receive events such as a sign-in, a submitted annotation and its time, or a recording made, under your account number — never your name, email or the content.

• Building an open Ukrainian construction-domain corpus of annotated texts and transcribed speech
• Research on speech recognition and named-entity recognition, and scientific publications about the corpus
• Releasing the corpus as an open dataset (Section 6)
• Running and securing the platform, and understanding how it is used so we can improve it

Your voice is used to build speech-recognition data — never to recognise who you are. Your speaker code exists only to keep your recordings together, for speaker-disjoint dataset splits and per-speaker statistics.

Your contributions and their open publication rest on your consent (Art. 11(1)(1) of Law No. 2297-VI; Art. 6(1)(a) GDPR), given by signing this document. Recording and publishing your voice rest on a separate, explicit consent — the voice box in the declaration, or the recorder’s question before your first take. If voice data are treated as special-category (biometric) data, this explicit consent is the condition we rely on (Art. 7(2)(1) of Law No. 2297-VI; Art. 9(2)(a) GDPR). Running your account, security and usage analytics rest on the steps you asked for and on our legitimate interests, as the Privacy Policy explains.

You may withdraw any consent at any time (Section 11); withdrawal does not affect the lawfulness of processing carried out before it.

From the corpus’s first public release — none has taken place yet — your contributions will be published as an OPEN dataset under the Creative Commons Attribution 4.0 (CC BY 4.0) licence on Hugging Face, with archival copies on Zenodo. By consenting, you agree that your annotations and recordings are released under that licence.

What a published recording carries: your voice, its transcript and the text you read, your speaker region, your speaker code, the day it was recorded and technical measures of the audio. Never published: your name, email address, town or village, your comments and notes, or your speaker profile answers.

Moderators can also choose individual approved recordings to be played on the website’s budov.org/voices page, with their text, region and recording day — never your name, speaker code or town.

⚠ This is pseudonymisation, not anonymisation: your voice remains your voice, and someone who knows it may recognise it. Once openly published, copies may be downloaded worldwide and CANNOT be recalled. Withdrawing consent keeps your data out of later releases (Section 11), but it cannot erase data that has already been released.

Record only your own voice. Do not record other people — no interviews, conversations, briefings or bystanders — even with their permission: another person cannot consent through you, and recordings with another voice are not accepted. A process for recordings with several speakers, each with their own consent, may come later; until then, a recording holds one voice, and it is yours.

Your data may be received by:

• The BUDOVA research team at KNUCA
• Our service providers, who process data on our behalf: Microsoft Azure (hosting, database, storage, email — EU region), GitHub (nightly backups), PostHog (analytics, EU cloud) and Hugging Face (dataset hosting — private until a release)
• The public — the open dataset, from its first release, freely downloadable worldwide; recordings chosen by the team on budov.org/voices

We store data in the EU; GitHub may process the backup copy in the United States. An open release means worldwide access by design. We do not sell personal data or share it with other organisations.

Your account, contributions and speaker profile are kept while you take part; recordings until you delete them, withdraw your voice consent or erase your account. Backups disappear within 30 days. The open dataset is permanent once released. After you erase your account we keep only proof of your consents — their type and version, how and when they were given or withdrawn, and a one-way hash of your email address — for 7 years after the end of the research project, so that we can show that data already published were contributed with consent.

Under Ukrainian data-protection law and, where it applies, the GDPR, you have the right to:

• Access your data
• Rectification of inaccurate data
• Erasure ("right to be forgotten")
• Restriction of processing
• Portability — export your data
• Objection to processing
• Withdraw consent at any time

How: the in-app GDPR panel (the gear icon, Settings → Your Data & Privacy) or email dolhopolov_sy@knuba.edu.ua. You may also lodge a complaint with the supervisory authority — the Ukrainian Parliament Commissioner for Human Rights — and, if you live in the EU or the EEA, with the data protection authority of your country.

You may withdraw at any time in the GDPR panel or by emailing dolhopolov_sy@knuba.edu.ua:

• This consent (data processing) — you can no longer contribute until you sign again, and your annotations stop counting: they are left out of the dataset from then on, and every task you annotated is recounted without you. Your account stays; deleting it is a separate step.
• Your voice consent — every recording of yours that is not yet in a public release is deleted, audio included. Recordings in a release already published cannot be recalled, but they are left out of every later release.

As Section 6 says, data already released openly cannot be recalled. When you erase your account, everything else is deleted too, except the proof of consent described in Section 9 and security-log entries in which a pseudonym replaces your email address.

At the consent gate you confirm:

• I have read and understood this information.
• I consent to the processing of my data for the stated purposes.
• I understand my contributions will be published openly (CC BY 4.0) and that this is irreversible.
• (optional) I consent to the recording of my own voice and its publication under a speaker code, as described above.
• I am 18 or older.

You sign by typing your full name.

• Your voice is described as it is handled: pseudonymised, not anonymised, and never used to recognise who is speaking.
• What a published recording carries — and what stays internal: your town, your notes, your speaker profile — is listed; recordings are stored without embedded file metadata; the /voices page is mentioned.
• Only your own voice may be recorded.
• Withdrawing does what it says: withdrawing this consent stops your annotations from counting, withdrawing the voice consent deletes your unpublished recordings.
• The voice consent is recorded with this document’s version, also when you tick the voice box at the gate.
• Product analytics (PostHog) and our service providers are named; no partner organisation receives your data.
• After erasure, only a minimal proof of consent is kept.
• Taking part is for adults (18+) only.
• Nothing has been published yet, so publication is described as what will happen.