SAM Privacy Notice

When you register interest, we store your email address, optional name, selected courses (or general SAM training updates), submission dates and a record of the permission you gave. We use your consent to send those training updates, including booking announcements. This does not subscribe you to unrelated marketing or create a SAM account.

You can withdraw using the private preference link shown after saving or included in training emails, or contact support@bigqualdatalabs.com. Withdrawal stops those updates; we keep the permission and withdrawal record so we can honour your choices. These records are available only through protected administration and are not included in public website content.

A training enquiry is handled as support correspondence and does not subscribe you to training updates. The provider, retention and rights information below also applies to these records.

BigQualDataLabs Ltd operates SAM (Social Analysis of Meaning). We are the controller for account, subscription, billing, support, security and service-administration information. For privacy questions or rights requests, contact support@bigqualdatalabs.com.

For customer research processed on documented instructions, we act as a processor. The researcher or institution is the controller: they decide the research purpose, permitted material, lawful basis and retention instructions. Account registration is not consent from research participants.

We receive account and billing details from you, workspace administrators and payment services; research files, text, images, metadata, coding and saved outputs from requested workflows; and support correspondence. We also keep linked sign-in, security, processing and AI-usage records for operating the service. Normal operational logs are intended to hold technical metadata rather than raw research content.

When you choose SAM Collect, we retrieve material from the selected public or official source. It can include authored text, names, handles, organisations, dates, URLs, source identifiers, engagement counts and media references. Public availability does not make information anonymous.

Account, access and relevant billing information is needed to provide the service. Without it, we may be unable to provide access or administer a subscription. Research inputs depend on the features you choose.

Contract: where you contract with us personally, we use the information necessary to provide your account, administer your subscription and deliver requested support (Article 6(1)(b)).

Legitimate interests: for institutional users, our interest is administering the institution's requested access, memberships and support. We also rely on legitimate interests to protect workspace boundaries, investigate technical failures or misuse, recover requested work and monitor service resource use (Article 6(1)(f)). This does not permit unrelated monitoring or reuse of research content.

Legal obligation: we use personal information necessary to meet applicable company accounting and tax-record duties (Article 6(1)(c)). This does not make every support or operational record a statutory record.

Customer research is processed to deliver, secure and support requested features, not for a separate general purpose of improving SAM. The research controller must establish its own lawful basis and, where relevant, a condition for special-category data and appropriate participant information.

Only upload or collect material you are entitled to process, with the required research, institutional, ethical, licensing or other authority. The public SAM route is for public, synthetic, de-identified low-risk or expressly approved pilot material. It is not a Safe Haven or NHS clinical system.

Do not upload identifiable patient or clinical data, highly sensitive special-category data, children's data, legal case data or other controlled data unless we have agreed and enabled a separate controlled-data route in writing.

YouTube Collect retrieves public text and metadata, including available transcripts and comments, not passwords, private messages or non-public account data. YouTube's own handling is described in Google's Privacy Policy.

Privacy Shield helps detect and redact common identifiers and selected identity fields. It is a review aid, not a guarantee of anonymisation. Review material and outputs in context, including what someone could infer by combining them with other information.

Requested AI features use OpenAI's API. SAM sends relevant research text or excerpts, selected images where applicable, analytical instructions and necessary context. These inputs and the generated outputs can contain personal information.

The production project is not opted into OpenAI's optional feedback, evaluation/fine-tuning or API input/output sharing. SAM also disables optional response storage. Standard abuse-monitoring logs can still contain inputs and outputs, generally for up to 30 days subject to OpenAI's exceptions. See OpenAI's API data controls.

The public service uses the standard API endpoint and does not promise UK-only or EU-only AI processing or Zero Data Retention. A separately agreed controlled-data route must specify its own provider, location and retention controls.

Supabase provides database, authentication and private file storage in London; Railway hosts the API and worker in Amsterdam; Vercel delivers the frontend globally with current server functions in the United States; and Upstash provides the queue, configured in London. Regional hosting does not locate every provider operation or support activity there. Provider information: Supabase, Railway, Vercel and Upstash.

OpenAI handles the AI inputs and outputs described above. Its DPA and subprocessor list describe its arrangements.

Stripe handles payments and billing, including some processing as a controller for fraud prevention and legal duties. SAM does not store full card details. Resend sends authentication and transactional emails and stores service data in the United States. Microsoft 365 handles support correspondence. Provider information: Stripe, Resend and Microsoft.

Bright Data, based in Israel, receives public searches, URLs and collection settings and handles YouTube results. These routes do not require private uploaded research files. Its privacy information describes its handling of account and public data. Other selected source services receive the searches or URLs needed for collection.

Cloudflare Turnstile receives verification tokens and browser/network signals for supported public forms. It acts as a processor for website protection and separately as a controller to improve bot detection, as described in its Turnstile Privacy Addendum.

Processing can take place outside the UK, including in the United States and Israel. The linked provider terms describe transfer provisions, including standard contractual clauses and the UK Addendum where applicable. Contact Support for details or a copy of applicable safeguards; a regional hosting choice alone does not establish them for every transfer.

Independent uploads and saved analyses remain with the workspace until an authorised deletion, an applicable source deadline or an agreed closure/deletion instruction. They do not have a general automatic 30-day expiry. Dataset deletion covers associated research records and registered private source files; active work must settle first and file-removal failures are reported. Shared or protected samples may require Support assistance.

Account closure, cancelling renewal and dataset deletion are separate. Contact Support to resolve ownership and linked research. Administrative archiving disables access and records a 30-day review window, but does not automatically erase research when that window ends.

For closed-account, support and operational records, retention depends on completing closure, support or recovery, specific continuing incidents or disputes, and applicable legal duties. These records are not all erased by dataset deletion. Billing records are retained for applicable accounting/tax duties and relevant transactions or disputes. Provider records follow their own applicable terms.

YouTube and Bluesky material collected through SAM Collect has the fixed 30-day availability period shown in SAM. It covers source text, identifiers, media references and copied quotations in derived outputs. Saving, analysing or making a team copy does not extend the deadline. Access ends at expiry even if removal is being retried. This rule does not apply generally to other Collect sources or independent uploads.

A separate, current researcher-reviewed summary may remain after source expiry following fidelity, contextual identification-risk and institutional-permission checks. Unreviewed or outdated drafts are removed. Such a summary does not preserve original source evidence or source-level reproducibility.

Deleted records can remain in Supabase's rolling seven-day daily database backups until those backups expire. Database backups exclude separately stored file contents. A restore must reapply the required deletion and expiry controls before access resumes. See Supabase's backup information.

SAM cannot remove copies downloaded by researchers or institutions. SAM's 30-day source rule does not set the retention of every provider-held copy. Contact Support about a deletion or retention enquiry.

Depending on the circumstances, you may have rights of access, correction, erasure, restriction and data portability. Contact support@bigqualdatalabs.com. For customer research, the research controller handles the substantive request and we assist as appropriate.

Your right to object: you can object to processing based on legitimate interests for reasons relating to your particular situation. Contact Support and we will consider your request under the applicable law.

You can complain to the Information Commissioner's Office, the UK supervisory authority.

We update this notice when relevant services, purposes or provider arrangements change and communicate material changes where appropriate.