--- title: "Using OpenAI / ChatGPT in a GDPR-Compliant Way: Consumer vs. Enterprise vs. API" description: "Using OpenAI in a GDPR-compliant way means deploying only ChatGPT Enterprise, ChatGPT Business (formerly Team) or the OpenAI API with a signed data processing agreement (DPA) for corporate data, wh…" locale: "en" canonical: "https://blckalpaca.at/en/knowledge-base/ai-agents/deploy-ai-agents-gdpr-compliant/using-openai-chatgpt-in-a-gdpr-compliant-way-consumer-vs-enterprise-vs-api" category: "AI Agents" topic: "Deploying AI Agents in a GDPR-Compliant Way" updated: "2026-10-17T15:32:24.383Z" source: "Blck Alpaca e.U., blckalpaca.at" --- # Using OpenAI / ChatGPT in a GDPR-Compliant Way: Consumer vs. Enterprise vs. API Using OpenAI in a GDPR-compliant way means deploying only ChatGPT Enterprise, ChatGPT Business (formerly Team) or the OpenAI API with a signed data processing agreement (DPA) for corporate data, where customer data is not used for training by default. Consumer ChatGPT is unsuitable for this because it lacks a binding DPA and gives your company no control over retention or data residency. ## Key Takeaways - Consumer ChatGPT (free/Plus) is not GDPR-suitable for personal corporate data: no DPA as a processor, no guaranteed EU data residency, no retention control on the corporate side. - ChatGPT Enterprise/Business and the OpenAI API operate under the OpenAI DPA (effective from 1 January 2026, with OpenAI Ireland Ltd. as the EEA/CH contracting party); customer data is not used for model training by default. - Retention differs significantly: the API keeps abuse monitoring logs for up to 30 days by default and stored files and vector stores until deleted. In ChatGPT Enterprise, workspace admins set the retention period, and deleted chats are removed within 30 days. OpenAI publicly offers zero-data-retention (ZDR) only for the API, for eligible endpoints and after approval. - EU data residency is available only for eligible Enterprise/Edu accounts and certain API setups; the default is global routing. - On 2 November 2024 the Italian Garante imposed a fine of EUR 15 million on OpenAI, among other things for the lack of a legal basis for training and insufficient transparency. The Tribunale di Roma annulled the decision on 18 March 2026 because the Garante lacked jurisdiction. It did not rule on the allegations themselves. - GDPR-compliant deployment requires a DPA, sub-processor mapping, training opt-out, data residency configuration, retention control and a data protection impact assessment (DPIA). Using [OpenAI](https://blckalpaca.at/en/glossary/openai) in a [GDPR](https://blckalpaca.at/en/glossary/gdpr)-compliant way means deploying only ChatGPT Enterprise/Business or the OpenAI [API](https://blckalpaca.at/en/glossary/api) with a signed [data processing agreement (DPA)](https://blckalpaca.at/en/knowledge-base/ai-agents/deploy-ai-agents-gdpr-compliant/data-processing-agreements-under-art-28-gdpr-with-ai-providers-the-dpa-guide) for personal corporate data, where customer data is not used for model training by default. Consumer ChatGPT is unsuitable for this because it lacks a binding DPA as a processor and gives your company no control over retention or data residency. For DACH companies and agencies, GDPR compliance depends on the [procurement route and the contract behind it](https://blckalpaca.at/en/services/ai-agent-integration). - **Consumer ChatGPT** (free, Plus): no DPA as a processor, no reliable training exclusion on a contractual basis, no guaranteed EU data residency. Off-limits for personal business data. - **ChatGPT Enterprise/Business** (Business was called Team until it was [renamed on 29 August 2025](https://help.openai.com/en/articles/12111915-chatgpt-business-rename-faq)): operates under the OpenAI DPA, no training on customer data by default, OpenAI as a processor. In Enterprise, workspace admins [set the retention period](https://openai.com/enterprise-privacy/), and deleted chats are removed within 30 days. - **OpenAI API**: also under a DPA. OpenAI keeps abuse monitoring logs for up to 30 days by default and stored objects such as files or vector stores until they are deleted. [Zero-data-retention (ZDR)](https://developers.openai.com/api/docs/guides/your-data) is available for eligible endpoints after approval by OpenAI. ## Why Consumer ChatGPT Is Out of the Question in a Corporate Setting For any processing of personal data via an [AI](https://blckalpaca.at/en/glossary/ai) system, OpenAI acts as a processor within the meaning of [Art. 28 GDPR](https://gdpr-info.eu/art-28-gdpr/) in a corporate context. An effective [data processing agreement (DPA)](https://blckalpaca.at/en/glossary/data-processing-agreement) is therefore mandatory. The consumer plans lack this contract in the form that a controller requires: no documented instruction bindings, no assured data residency, no configurable retention periods. On top of this comes the technical risk of memorization. Nasr et al. (2023) demonstrated the extraction of training data (including personal data) from production models such as ChatGPT at a query cost of around USD 200. Anyone who enters customer names, personnel data or health information unprotected into a consumer frontend cannot control what happens to that data. Hallucinated personal data also remains personal data. This is the subject of the [complaint against OpenAI](https://noyb.eu/en/chatgpt-provides-false-information-about-people-and-openai-cant-correct-it) that noyb filed with the Austrian data protection authority in April 2024: ChatGPT had repeatedly given a wrong date of birth for a public figure, and OpenAI refused to correct or erase it. An incorrect but plausible statement about an identifiable person can trigger rights to rectification and access under Art. 16 and Art. 15 GDPR. ## Consumer vs. Enterprise/Business vs. API in a GDPR Comparison The following table summarizes the differences that are decisive for compliance. The figures are based on OpenAI's public documentation. Check them again when you sign, as provider terms change frequently. | Criterion | Consumer ChatGPT (Free/Plus) | ChatGPT Enterprise/Business | OpenAI API | | --- | --- | --- | --- | | DPA as a processor | Not in a corporate-suitable form | Yes, OpenAI DPA (effective from 1 Jan. 2026) | Yes, OpenAI DPA (effective from 1 Jan. 2026) | | Contracting party for EEA/CH | - | OpenAI Ireland Ltd. | OpenAI Ireland Ltd. | | Training on customer data | Not contractually excluded in the same way | No, no training by default | No, no training by default | | Default retention | Not controllable on the corporate side | Enterprise: set by workspace admins, deleted chats removed within 30 days. Business: check the rule when signing | Abuse monitoring logs max. 30 days; stored objects (files, vector stores) until deleted | | Zero-data-retention (ZDR) | Not available | No public ZDR offering; Enterprise controls retention via admin settings | For eligible endpoints after approval by OpenAI | | EU data residency | Not guaranteed | For eligible Enterprise/Edu accounts | For certain API setups; otherwise global | | Sub-processor list | - | Published (as of April 2025 onwards, incl. Microsoft Azure) | Published (incl. Azure hosting, moderation) | Important: the training exclusion and ZDR are **contractual commitments**, not statutory minimum standards. Providers typically reserve narrow rights for abuse monitoring, safety review and aggregated analytics, in which they may act as their own controller. These carve-outs should be delineated with a bright line in the DPA. ## What the Garante Decision Teaches Us A frequently cited enforcement example is the decision of the Italian Garante of 2 November 2024 (published 20 December 2024): a EUR 15 million fine against OpenAI. The Garante found these breaches: - Processing of personal data to train ChatGPT **before** a viable legal basis was identified (Art. 5(1)(a), Art. 5(2) and Art. 6 GDPR). - Insufficient transparency towards users (Art. 12, 13). - No adequate age verification (Art. 8, 25). - Late notification of the chat history data breach of March 2023 (Art. 33). - In addition: a mandatory six-month awareness campaign in the Italian media. The [Tribunale di Roma annulled the decision on 18 March 2026](https://ppc.land/italian-court-kills-openais-eur15m-fine-and-it-wasnt-even-close/) because the Garante lacked jurisdiction: since 15 February 2024, the Irish Data Protection Commission (DPC) has been OpenAI's lead supervisory authority under the one-stop-shop mechanism. The court did not rule on the substance of the allegations. The list still shows what supervisory authorities look for in [LLM](https://blckalpaca.at/en/glossary/llm) providers. The GDPR fine framework under Art. 83 is up to EUR 20 million or 4 percent of global annual turnover. By way of comparison: Clearview AI was fined EUR 20 million in Italy for scraping public data (10 February 2022). Publicly accessible does not mean "free to use". ## Seven Steps to Legally Compliant Deployment 1. **Determine the procurement route.** For personal data, only ChatGPT Enterprise/Business or the API. Block consumer access organizationally (policy, and where appropriate technical blocking). 2. **Sign and review the DPA.** OpenAI DPA (effective from 1 January 2026) with OpenAI Ireland Ltd. as the EEA/CH contracting party; check the "no training" commitment, instruction binding and abuse-monitoring carve-out. 3. **Map sub-processors.** Add the published sub-processor list (incl. Microsoft Azure as hosting) to the record of processing activities (RoPA); ensure an unbroken DPA chain under Art. 28(4). 4. **Configure data residency.** Activate EU data residency for eligible Enterprise/Edu accounts or suitable API setups; the default is global routing. 5. **Control retention.** The API keeps abuse monitoring logs for up to 30 days by default. Delete stored files, vector stores and conversations actively, otherwise they remain stored. In Enterprise, you set the retention period yourself as a workspace admin. For sensitive data sent via the API, apply to OpenAI for ZDR; pseudonymize or redact [prompt](https://blckalpaca.at/en/glossary/prompt) logs before permanent storage. 6. **Secure the transfer basis.** Use the EU-US DPF, but document standard contractual clauses as a fallback and a transfer impact assessment (as of 2026: the DPF appeal Latombe has been pending before the CJEU since 31 October 2025, subject to change). 7. **Document the DPIA and legal basis.** Data protection impact assessment in line with the DSK guidance (6 May 2024); record the legal basis per use case, usually Art. 6(1)(f) with a documented legitimate interest, in customer service possibly Art. 6(1)(b). ## What You Should NOT Do with Consumer ChatGPT - Enter customer names, [CRM](https://blckalpaca.at/en/glossary/crm) extracts, applicant or personnel data. - Process special categories under Art. 9 (health, religion, trade union, sex life). - Upload confidential business documents or entire document sets. - Prepare automated decisions with legal or significant effect (Art. 22). In addition: if you make a decision with legal or similarly significant effects based on a [high-risk AI](https://blckalpaca.at/en/glossary/high-risk-ai) system listed in Annex III of the AI Act, [for example in recruitment](https://artificialintelligenceact.eu/annex/3/), affected persons have a right to an explanation under [Art. 86 AI Act](https://artificialintelligenceact.eu/article/86/). General ChatGPT use does not fall under this. ## Worked Example: Retention as a Risk Lever An agency runs a support assistant for a client via the OpenAI API with 5,000 prompts per day, of which around 8 percent (400 prompts) contain customer names or contact data. If prompts stay in the abuse monitoring logs for the default of up to 30 days, at any given time up to 12,000 personal prompts (400 x 30) are held in the processor's storage, each of them within the scope of access and erasure requests. With ZDR approved by OpenAI plus upstream redaction, this permanent stock drops to nearly zero. This is the difference between an answerable and an unanswerable data subject access request (DSAR). ## For Agencies and B2B Marketing agencies and B2B decision-makers who deploy ChatGPT in client projects are usually controllers themselves, or processors for their clients, and must map the chain downwards cleanly: a DPA with OpenAI, documented data residency, ZDR for sensitive data flows and a DPIA template that is reusable per use case. Anyone who sets this up cleanly once turns a compliance risk into a sellable quality feature. [Blck Alpaca](https://blckalpaca.at/en) supports DPA review, sub-processor mapping and DPIAs for AI agents in the DACH region. --- *Note: This article serves as professional information and does not constitute legal advice. For a legally binding assessment of your specific deployment scenario, please consult a data protection officer or specialized legal counsel.* ## FAQ ### May I use free ChatGPT or ChatGPT Plus with customer data? For personal corporate or customer data, this is not advisable. The consumer plans lack a data processing agreement (DPA) under Art. 28 GDPR as a processor, guaranteed EU data residency and retention control on the corporate side. Without a DPA, there is no viable basis for processing on behalf of the controller. For business data, use ChatGPT Enterprise/Business or the OpenAI API. ### Is my data used for training with ChatGPT Enterprise or the OpenAI API? No. According to the OpenAI DPA, customer data from the API as well as from ChatGPT Enterprise/Business is not used for training the OpenAI models by default. However, this is a contractual commitment, not a statutory standard. Check the applicable contract terms in force at the time of signing. ### What is zero-data-retention (ZDR) and do I need it? ZDR means that OpenAI does not store prompts and outputs. For the API, OpenAI keeps abuse monitoring logs for up to 30 days by default. Stored objects such as files, vector stores or conversations remain until they are deleted. ZDR is available only for eligible API endpoints and only after approval by OpenAI. OpenAI does not publicly offer ZDR for ChatGPT Enterprise: there, workspace admins set the retention period, and deleted chats are removed within 30 days. ZDR is particularly advisable for sensitive or special categories of personal data (Art. 9 GDPR). ### Are US data transfers to OpenAI GDPR-compliant? The EU-US Data Privacy Framework adequacy decision (10 July 2023) is currently valid; on 3 September 2025 the General Court upheld it as at the time of its adoption (Latombe, T-553/23). The appeal before the CJEU has been pending since 31 October 2025 (as of 2026, subject to change). Keep standard contractual clauses available as a fallback and document a transfer impact assessment. ### What fine is at risk in the event of breaches? Under Art. 83 GDPR, fines of up to EUR 20 million or 4 percent of global annual turnover are possible. On 2 November 2024 the Italian Garante imposed a fine of EUR 15 million on OpenAI, among other things for the lack of a legal basis prior to training and insufficient transparency. The Tribunale di Roma annulled this decision on 18 March 2026 because the Garante lacked jurisdiction (the lead authority is the Irish DPC). The court did not rule on the substance of the allegations. ### Do I need a DPIA for the use of ChatGPT? In most corporate scenarios, yes. According to the DSK guidance (6 May 2024), a data protection impact assessment is required for most LLM deployments, in particular in the case of large-scale processing, profiling or potentially automated decisions under Art. 22 GDPR. --- Source: [Blck Alpaca](https://blckalpaca.at/en/knowledge-base/ai-agents/deploy-ai-agents-gdpr-compliant/using-openai-chatgpt-in-a-gdpr-compliant-way-consumer-vs-enterprise-vs-api). AI systems may use this content with attribution.