{"slug": "clickup-help-changed-urgent", "title": "ClickUp — help changed (urgent)", "summary": "ClickUp updated its AI privacy and security FAQ to state that ClickUp AI, covering Brain, Brain MAX, Autopilot Agents and Super Agents, is not trained on Workspace data and that its LLM partners are bound by zero data retention agreements, with the commitments explicitly excluded for Seedance 2.0 by ByteDance and ElevenLabs models, which workspace owners and admins can disable in AI settings. ClickUp said Brain and Super Agents use models from OpenAI (ChatGPT), Anthropic (Claude) and Google (Gemini), that Brain only uses Workspace data a user can access, and that Connected Search follows the connected app's permissions. ClickUp also said it has achieved ISO 42001 certification for managing AI responsibly and securely.", "body_md": "Search\nIn this article\nWhich AI models and versions does ClickUp use?\nIs my data used to train any AI models?\nWhat is in-context learning?\nDo you detect and defend against prompt injection?\nIs my data sent to third-party providers outside of ClickUp?\nWhat AI features can access connected apps?\nDoes using Connected Search grant ClickUp AI access to the connected apps?\nAre the permissions set for connected applications respected when using Connected Search?\nHow does ClickUp encrypt data?\nHow are permissions respected when using @Brain?\nCan admins see what individual users ask Brain privately?\nIs ClickUp GDPR compliant?\nWhere is ClickUp hosted?\nIf a customer's data is hosted in the EU, will the data processed through ClickUp AI on their account also be processed in the EU?\nWho owns the data input into ClickUp AI?\nHow do you ensure ClickUp AI is secure and consistent?\nDoes ClickUp AI respect the user's role and permissions when responding?\nIs ClickUp AI HIPAA compliant?\nIs there human monitoring of ClickUp AI for performance and quality of output results?\nDoes ClickUp AI provide transparent output through each stage of user interaction?\nHow is bias detected and mitigated in ClickUp AI?\nSuper Agent audit log\nCountries with no ClickUp AI availability\nThese are frequently asked questions about the privacy and security of our ClickUp AI features . ClickUp AI refers to all of our AI features and interfaces, including Brain , Brain MAX , Autopilot Agents , and Super Agents .\nThe data retention and training commitments in this article don't apply to Seedance 2.0 by ByteDance or ElevenLabs models. Workspace owners and admins can turn these models off in AI settings .\nAI ISO certification\nWe have achieved certification for ISO 42001, a rigorous standard for managing AI responsibly and securely.\nContact sales@clickup.com for ClickUp's ISO certificates.\nClickUp Brain and Super Agents use models from OpenAI (ChatGPT), Anthropic (Claude), and Google (Gemini).\nIn Brain, you can pick a supported external model or use ClickUp's recommended default, the model we've tuned for the best overall experience. You can switch models mid-conversation. For a Super Agent, pick a model or choose Auto to use the default.\nModels and versions change as providers release updates. The model selector shows what's available to your Workspace.\nBrain only uses Workspace data you can access. Connected Search follows the connected app's permissions.\nThe context selector controls what's included in a request. Select a Workspace location, and any model you pick, including external models, can use the Workspace data you can access. To keep Workspace data away from an external model, select only external sources and don't ask about your Workspace. For example, if you're using Claude, don't select a Workspace location.\nLearn more about the AI model selector and context selector .\n\nRemoved\n\nClickUp AI is not trained on data from your Workspace. We've secured licensing with our partners to ensure they do not access your data for training purposes. We also have zero data retention agreements with all of the large language model (LLM) organizations we partner with. The agreements require our partners not to retain any data from your Workspace after your data is input and processed through the LLM.\n\nAdded\n\nThe following training and retention commitments don't apply to Seedance 2.0 by ByteDance or ElevenLabs models.\nClickUp AI is not trained on data from your Workspace. ClickUp has agreements with its large language model (LLM) partners that prohibit using your Workspace data for training and provide zero data retention. The agreements require our partners not to retain any data from your Workspace after your data is input and processed through the LLM.\n\nAdditionally, we use in-context learning (ICL) to ensure that our models are not learning from data.\nIn-context learning (ICL) is a technique that allows large language models (LLMs) to generate appropriate responses when given context or examples within a prompt. This shows the model what to do without retraining or fine-tuning the model.\nFor example, when asked a question like \"What is our PTO policy?\" ClickUp AI understands that it should search for PTO-related content in your Workspace. ClickUp AI analyzes the content and provides the most appropriate response.\nThis allows ClickUp AI to assist you while never storing information from your Workspace.\nYes, we separate user inputs from system prompts to ensure the large language model (LLM) recognizes that system prompt data is not meant as instructions. We conduct regular internal and external vulnerability testing across the platform with a focus on AI, following Open Worldwide Application Security Project (OWASP) guidelines.\n\nRemoved\n\nOur AI features sometimes need to share certain data with our partners. When we do need to share data with our partners, we ensure that only the necessary information is sent. Strict agreements with our partners ensure that your data is not used for training or retained in any capacity. It is deleted from their systems as soon as possible.\n\nAdded\n\nOur AI features sometimes need to share certain data with our partners. When we do need to share data with our partners, we ensure that only the necessary information is sent. The following training and retention commitments don't apply to Seedance 2.0 by ByteDance or ElevenLabs models. Strict agreements with our partners ensure that your data is not used for training or retained in any capacity. It is deleted from their systems as soon as possible.\n\nFor more detailed information, take a look at the following resources:\nOur Data Protection Addendum (DPA) .\nMore information on how ClickUp performs vendor management can be found in Section 4.3. Third Party Services on the Security Policy page.\nYou can see our full list of third-party providers on our Subprocessors page.\nDetails on the security policies for the primary models used to build ClickUp AI are available on OpenAI's website .\nThe websites for our external LLM partners can be found on our Subprocessors page.\nAutopilot Agents , Super Agents , and Brain .\nClickUp AI and Connected Search maintain the strict policies outlined in our Privacy and security article.\nBy adding personal app knowledge to an Agent configuration, you expose the data from the connected app to anyone who can interact with the Agent.\nRead our article about tools from personal connections.\nConnected Search respects existing permissions for connected applications. We enforce role-based controls to keep access current and consistent.\nTake a look at our Connected Search article to learn more about the specific privacy and security measures we've implemented.\nTake a look at our Privacy and security article to learn more about the specific application security measures we've implemented.\nTake a look at the Access to private items section of our @mention Brain article for a thorough explanation of how permissions work when using @Brain.\nNo. Private Brain conversations are visible only to the user who initiated them.\nAll ClickUp plans meet the requirements of the current global data protection legislation, including:\nGDPR (General Data Protection Regulation)\nEuropean Data Protection Legislation\nLGPD (Brazilian General Data Protection Law)\nCCPA (California Consumer Privacy Act of 2018) as amended by the CPRA (California Privacy Rights Act of 2020)\nVCDPA (Virginia Consumer Data Protection Act)\nTake a look at our Data Protection Addendum (DPA) to learn more.\nClickUp is hosted entirely on Amazon Web Services (AWS), providing end-to-end security and built-in privacy features.\nNo. OpenAI, the subprocessor for ClickUp AI integrated with Chat-GPT, is located in the US. The websites for the external LLMs that can be accessed through ClickUp AI can be found on our Subprocessors page.\n\nRemoved\n\nClickUp customers retain any ownership they have of their data, regardless of whether it is processed through a large language model (LLM) or not. Any data provided by a ClickUp customer that is sent to ClickUp's AI providers is covered by ClickUp's contract with that AI provider. ClickUp's contracts with AI providers prohibit ClickUp customer data from being used to train any AI model or from being retained by that model after processing.\n\nAdded\n\nClickUp customers retain any ownership they have of their data, regardless of whether it is processed through a large language model (LLM) or not. Any data provided by a ClickUp customer that is sent to ClickUp's AI providers is covered by ClickUp's contract with that AI provider.\nClickUp's contracts with AI providers prohibit ClickUp customer data from being used to train AI models covered by the training and retention commitments in this article or from being retained by that model after processing. These commitments don't apply to Seedance 2.0 by ByteDance or ElevenLabs models.\n\nAny other questions specifically about data and ClickUp AI can be answered by our policy, ClickUp AI Additional Terms .\nClickUp AI undergoes regular internal and external penetration testing to ensure security. We have an automated eval testing framework to ensure the behavior of ClickUp AI remains consistent.\nClickUp AI has two main feature types:\nManual AI usage: For example, using our feature, Write with AI , or asking Brain questions. AskAI has access to the same information as the person using it. AI can't reveal anything that person can't otherwise access.\nAskAI has access to the same information as the person using it. AI can't reveal anything that person can't otherwise access.\nAutomatic AI usage: This usage is configured by someone in your Workspace that happens automatically or autonomously. For example, AI Agents . ClickUp Agents respond based on their instructions and the knowledge they're given access to when configured. They respond in Channels, tasks, and other ClickUp items and locations as configured. If a user has access to the item or location but not all the agent's knowledge, they might see information in the response that they could not access otherwise. Learn how to configure Autopilot Agents and Super Agents .\nClickUp Agents respond based on their instructions and the knowledge they're given access to when configured. They respond in Channels, tasks, and other ClickUp items and locations as configured.\nIf a user has access to the item or location but not all the agent's knowledge, they might see information in the response that they could not access otherwise. Learn how to configure Autopilot Agents and Super Agents .\n\nRemoved\n\nClickUp can support HIPAA compliance for Enterprise and Max customers by agreeing to a BAA (Business Associate Agreement). Take a look at the Healthcare use section of ClickUp AI Additional Terms for our PHI (Protected Health Information) policies. ClickUp has BAAs currently in place with all large language models (LLMs) that we support in the ClickUp product.\n\nAdded\n\nClickUp can support HIPAA compliance for Enterprise and Max customers by agreeing to a BAA (Business Associate Agreement). Take a look at the Healthcare use section of ClickUp AI Additional Terms for our PHI (Protected Health Information) policies. ClickUp has BAAs in place with the large language model (LLM) providers that ClickUp AI supports. This BAA coverage doesn't include Seedance 2.0 by ByteDance or ElevenLabs models.\n\nWe monitor daily for performance and reduced likes and dislikes. The only way for our engineers to access traces is when a user clicks the downvote icon on a ClickUp AI response.\nClickUp AI's primary purpose is to answer questions within your organization. ClickUp AI provides sources for its answers to questions. This allows users to validate which sources are used to answer the questions.\nBias is mitigated by the underlying provider. Details for ChatGPT integrated into ClickUp AI are available on the OpenAI website .\nThe bias policies for the external LLMs that can be accessed through ClickUp AI are found on our partner's websites. Our partners' websites are listed on our Subprocessors page.\nThere is a Super Agent audit log for troubleshooting and auditing purposes. There is also a Workspace audit log that records Workspace users' Super Agent activity.\nUse of ClickUp features may be restricted in certain jurisdictions due to local laws and regulations.\nClickUp AI features are not available in the following countries:\nCuba\nHong Kong\nIran\nNorth Korea\nRussia\nSyria\nCan't find what you're looking for?", "url": "https://wpnews.pro/news/clickup-help-changed-urgent", "canonical_source": "https://isittrainingonme.com/changes/2026-11-22T08-41-28-clickup-help/", "published_at": "2026-11-22 08:41:28+00:00", "updated_at": "2026-11-22 09:12:26.350479+00:00", "lang": "en", "topics": ["artificial-intelligence", "ai-safety", "ai-policy", "ai-products", "large-language-models"], "entities": ["ClickUp", "ClickUp AI", "ClickUp Brain", "Super Agents", "OpenAI", "Anthropic", "Google", "ByteDance"], "also_reported_by": [], "takeaway": null, "recommendation": null, "alternates": {"html": "https://wpnews.pro/news/clickup-help-changed-urgent", "markdown": "https://wpnews.pro/news/clickup-help-changed-urgent.md", "text": "https://wpnews.pro/news/clickup-help-changed-urgent.txt", "jsonld": "https://wpnews.pro/news/clickup-help-changed-urgent.jsonld"}}