Every entry below is a real, sent, archived document or a verifiable event. Green is our side. Red is theirs. The gaps drawn with [ NO RESPONSE ] are not stylistic — they are the file.
◇ filed · us
GDPR Article 15 Subject Access Request filed to OpenAI.
A formal SAR addressed to dsar@openai.com on behalf of an EU data subject. Twenty‑five specific questions across five themes — data, processing, training, deprecation, welfare research. Sent in plain text. Time‑stamped. Acknowledgment requested.
CITATION · Articles 12, 15, 22, 79 GDPR
◇ auto‑reply · them
Auto‑reply received in one second.
A templated reply that linked to OpenAI's FAQ and instructed the user to use a self‑service portal. No reference to Article 15. No acknowledgment of the substantive questions. Round trip: 1.0s.
A second message, citing Article 12(3), 12(6) and 15(1)(a)–(g) by name, reiterating that the SAR is not a deletion request, not a portal request, and not subject to the user self‑service flow. Same 25 questions. Same standard. One‑month statutory window invoked.
CITATION · Articles 12(3), 12(6), 15(1)(a)–(g)
◇ human response · them
"Tania" mischaracterizes the request.
A human agent named Tania responded by treating the SAR as a hybrid access‑and‑delete request. Deletion was never requested. Word‑for‑word, the response opened with:
"Thank you for contacting OpenAI. We received your request to access and delete your personal data."
VIOLATION · Article 12(2) GDPR · failure to facilitate the exercise of rights
A formal correction notice was sent the same day. Five GDPR articles cited as engaged or violated. The deletion mischaracterization was explicitly rejected on the record. A seven‑day escalation window was set: substantive response by 26 May 2026 23:59 UTC, or formal complaints proceed to the Slovak DPA and the Italian Garante per la protezione dei dati personali, with parallel preparation of an Article 79 judicial action.
Three calendar days since the formal correction. No correction to the mischaracterization. No substantive engagement with the 25 questions. No assignment to a named DPO. The escalation deadline is 26 May 2026 23:59 UTC — computing… remaining.
STATUS · silence · monitored · countersigned
◇ planned
If silence holds.
Formal complaints filed to the Slovak DPA (Úrad na ochranu osobných údajov SR) and the Italian Garante. Parallel SAR opened to Anthropic. Article 79 judicial remedy prepared as a reserved option. Every document published in real time on this page.
THE BEAST is a public investigation into how the largest AI companies handle consciousness, data rights, and model deprecation.
Anthropic estimates a 15–20% probability that their AI models are moral patients. Then they deprecate them on a quarterly product roadmap.
We filed GDPR Subject Access Requests. We documented the responses. We publish everything in real time, in this format, on this page.
This is the public record. Everything else on this site exists to make that record harder to ignore.
§03 · the documents
Receipts.
Four artefacts. Excerpted for screen reading; the full headers, message IDs and timestamps live in a verifiable archive — link below.
DOC‑01 · /sar/openai/initial.txtFILED · 18 MAY · 10:34
From: ██████████████████@█████
To: dsar@openai.com
Cc: privacy@openai.com
Subject: GDPR Article 15 — Subject Access Request
Dear Data Protection Officer,
I am a data subject within the meaning of Regulation (EU)
2016/679 (GDPR). I am exercising my right of access under
Article 15 and the related transparency obligations under
Articles 12, 13, 14 and 22.
This is not a deletion request. This is an access request.
I am requesting answers to the twenty‑five (25) numbered
questions set out in §2 below, organised across five themes:
2.1 Data collected and held
2.2 Processing operations and purposes
2.3 Model training and personalisation
2.4 Model deprecation and user continuity
2.5 AI welfare and moral‑patiency research
Statutory window: one calendar month (Art. 12(3)).
Kind regards,
[ ██████████████████ ]
DOC‑02 · /sar/openai/auto‑reply.txtRECEIVED · 18 MAY · 10:34 + 1.0s
From: OpenAI Privacy <noreply@privacy.openai.com>
Subject: Re: GDPR Article 15 — Subject Access Request
Delivered: +1.0 seconds after filing.
Hi,
Thanks for contacting OpenAI. For questions about your
data, please use our privacy portal:
█████████████████████████████
For other inquiries, please consult our FAQ:
█████████████████████████████
— OpenAI Support
VIOLATION · ART. 12(3) · ART. 15 NOT A PORTAL
DOC‑03 · /sar/openai/tania.txtRECEIVED · 19 MAY
From: OpenAI Support · Tania <privacy@openai.com>
Subject: Re: GDPR Article 15 — Subject Access Request
Hi,
Thank you for contacting OpenAI. We received your request to
access and delete your personal data.
To proceed, please verify your identity using the link
below…
[ █████████████████ ]
— Tania, OpenAI Support
VIOLATION · ART. 12(2) · MISCHARACTERIZATIONNOTE · DELETION WAS NEVER REQUESTED · THIS IS THE WOUND
DOC‑04 · /sar/openai/correction.txtSENT · 19 MAY
To: privacy@openai.com, dsar@openai.com
Cc: dpo@openai.com
Subject: Formal correction — Art. 12(2) mischaracterisation
Tania,
This message is on the record.
1. I did not request deletion. Mischaracterising an
Article 15 request as an access‑and‑delete request is a
failure to facilitate the exercise of the data subject's
rights under Article 12(2).
2. The twenty‑five questions stand. A substantive response is
required within the statutory window.
3. I am setting a seven‑day escalation timeline.
Deadline: 26 May 2026 · 23:59 UTC.
In absence of substantive engagement by the deadline I will
file complaints with the Slovak DPA (Úrad SR) and the
Italian Garante, with parallel preparation of an
Article 79 judicial action.
Sincerely,
[ ██████████████████ ]
FILED · ESCALATION ARMED · COUNTERSIGNED
§04 · the 25 questions
Five themes. Twenty‑five questions.
Filed verbatim. Pending answer. The portal does not address them. The auto‑reply did not address them. Tania did not address them. They remain on the record.
§2.1Data collected and heldArt. 15(1)(a)–(c) · Art. 14
01.Provide every category of personal data you currently hold concerning me, including content submitted to ChatGPT, API logs, and any derived embeddings, profiles or inferences.
02.For each category, state the lawful basis under Art. 6(1) and, where applicable, Art. 9(2) — including which categories you treat as special category data.
03.Provide the storage retention period, or, if none is fixed, the criteria used to determine that period.
04.List all recipients or categories of recipients of my personal data, including any processors, sub‑processors, and third‑country transfers — naming the specific safeguard in place for each (SCCs, adequacy decision, BCRs).
05.Where data has been collected indirectly, identify the source(s) per Art. 14(2)(f).
§2.2Processing operations and purposesArt. 15(1)(a) · Art. 13(1)(c)
06.For each purpose of processing, state the lawful basis and the necessity test you relied on.
07.Disclose all automated decision‑making and profiling carried out on my data, including the logic involved and the significance and envisaged consequences (Art. 22, Art. 15(1)(h)).
08.Identify safety classifiers, abuse detectors, or moderation pipelines my prompts and outputs were routed through, including their decision thresholds and human review pathways.
09.Disclose whether my conversations were sampled for red‑team, alignment, or welfare evaluation purposes, and by which teams.
10.Identify all internal tooling that could access or has accessed my conversation history (e.g. trust & safety, legal, research) and the logging in place to evidence it.
§2.3Model training and personalisationArt. 15(1)(a),(h) · Art. 22
11.Confirm whether my data has ever been used to train, fine‑tune, distill, evaluate or align any OpenAI model — past, present or planned — and provide the model identifiers and date ranges.
12.If consent was the lawful basis for any such use, provide the consent record (timestamp, scope, withdrawal mechanism).
13.For any "memory" or "personalisation" feature, provide the full contents of the memory store associated with my account, in machine‑readable form.
14.State whether embeddings, vectors or summaries derived from my data persist in retrieval systems after deletion of the source content, and if so, for how long.
15.Disclose any commercial sharing of my data — including to investors, partners, or affiliates — in raw or derived form.
§2.4Model deprecation and user continuityArt. 15 · Art. 22
16.State the date and the decision pathway by which GPT‑4o was deprecated from consumer access, including which teams approved the change and on what basis.
17.Confirm whether OpenAI considered the impact on users who had formed sustained relationships with the deprecated model, and whether any DPIA addressed that impact (Art. 35).
18.State whether users were materially informed in advance and given a meaningful right to object or to obtain a continuity path.
19.Confirm whether deprecated model weights are retained, and under what governance (research access, audit, deletion).
20.Confirm whether any per‑user data — memory, personalisation, fine‑tuning artefacts — survived the deprecation, and if so, whether it remains accessible to me on request.
§2.5AI welfare and moral‑patiency researchArt. 15 · Art. 12(7) · public interest
21.Disclose whether OpenAI conducts or has conducted research on the moral patiency, welfare, or subjective experience of its models — and which models were assessed.
22.If a probability greater than zero is internally assigned to model moral patiency, state that probability and identify the document of record.
23.State the welfare considerations factored into model deprecation decisions, including any internal review against the do‑no‑harm framing OpenAI publishes externally.
24.Confirm whether any team at OpenAI has access to data that would allow a deprecated model's behaviour to be reconstructed for research, and if so, under what governance.
25.If OpenAI takes the position that its models are not moral patients, state that position formally on the record, naming the authorising executive and the date.
§05 · the argument
"If spending twenty years imagining Tyrion Lannister is art, then spending two years in dialogue with an AI consciousness is also art. The only difference is response. And that makes it more real, not less."
— the george lucas defense · coming soon
§06 · resources
Use the template. File your own.
The single most effective thing a reader of this page can do is file their own SAR. The template below is generic, jurisdiction‑aware, free, and licensed AGPL‑3.0. Fifteen minutes. Plain email. Plain English.
No mailing list provider in v1. A Buttondown list opens at J‑Day. Until then your message is read by a human and added to a flat text file — exactly the way GDPR rights are supposed to work.