DeepSeek
Privacy policy rating
- Your data trains their AI unless you opt out by form or email.
- The policy rules out using your data for targeted ads.
- Data is shared only with the providers who run the service.
- The documents do not say who can read your content.
Chats can train models unless you email to opt out; data is stored in China, and developer terms cap and let them close accounts.
Agent Trust Hub uses AI and may make mistakes. Review reports and confirm their contents before relying on them.
- Capped at D (the answers alone would give B): Exposed database or plaintext chats in the last 24 months · Publicly reachable ClickHouse database with over a million log lines including chat history (found by Wiz Research) (2025)
Collected
Account info, chats, uploads, voice and photos, device IDs, IP, and approximate location.
Read the exact line
Privacy policy
“we may collect your text input, voice input, prompt, uploaded files, photos, feedback, chat history, or other content that you provide to our model and Services”
Training
They use your data to run the service and to train and improve their AI models.
Read the exact line
Privacy policy
“To improve and develop the Services and to train and improve our technology, such as our machine learning models and algorithms.”
Sharing
Service providers and their company group can see it; they say they do not sell it.
Read the exact line
Privacy policy
“We share Personal Data You Provide, Automatically Collected Personal Data, and Personal Data From Other Sources with these service providers as necessary”
Kept
They keep account and chat data while you have an account, and longer if they say they need it.
Read the exact line
Privacy policy
“when we process your Personal Data to provide you with the Services, we keep this Personal Data for as long as you have an account.”
Controls
You can delete chats in settings and email them to opt out of training or .
Read the exact line
Privacy policy
“the right to opt-out of using your Personal Data for training our models or optimizing our technologies.”
Fine print
Open-platform terms are as-is, you them, and they can close accounts without prior notice.
Read the exact line
Terms of service
“DeepSeek reserves the right to independently judge and take measures against you, including issuing warnings, setting deadlines for correction, restricting account functions, restricting or suspending usage, locking or closing accounts”
Expand “Read the exact line” to see the source alongside the explanation.
What you can turn off
The controls and opt-outs their own documents describe, and where they say to find them.
Email to opt out of model training
privacy@deepseek.com
Delete or copy chat history
In-app settings
Request access, correction, or deletion
privacy@deepseek.com (EEA/UK also rep_deepseek@prighter.com)
Delete your account if you want to leave
Account settings; they say you cannot reactivate or retrieve content after deletion
If a switch is not where they say, the deletion request above still applies.
Line by line
The lines that matter most, worst first.
What you type and upload can be used to train their AI unless you opt out.
Read the exact line
Privacy policy
“To improve and develop the Services and to train and improve our technology, such as our machine learning models and algorithms.”
For example, a private story you paste into the chatbot could be used to improve future models.
Your personal data is stored in China, even if you live elsewhere.
Read the exact line
Privacy policy
“To provide you with our services, we directly collect, process and store your Personal Data in People's Republic of China.”
For example, chats from a user in Europe are processed on servers in China.
They can lock or close a developer account without warning if they think you broke the rules.
Read the exact line
Terms of service
“without the need for prior notification, if DeepSeek believes or determines that: (a) you and/or your end user may be in breach of these Terms or any applicable law;”
For example, they could shut off your API key overnight and you would have no notice first.
If someone sues them because of how you used the service, you may have to pay their legal bills.
Read the exact line
Terms of service
“You agree to , defend, and hold us and our affiliates and licensors (if any) harmless against any liabilities, damages, and costs (including reasonable attorneys' fees)”
For example, if an end user sues over an app you built on their API, you might have to cover DeepSeek’s lawyer fees.
They do not promise the service will work, be accurate, or stay online.
Read the exact line
Terms of service
“THE SERVICES ARE PROVIDED ON AN “AS IS” AND “AS AVAILABLE” BASIS AND WE MAKE , REPRESENTATION OR CONDITION TO YOU WITH RESPECT TO THEM”
For example, if the API is down and your product fails, they say they are not responsible.
They may scrape public web data about people to train models.
Read the exact line
Privacy policy
“We may obtain publicly available Personal Data via online sources to train our models and provide Services.”
For example, a blog post with your name could be used as training data.
You can ask them not to use your personal data for training, by contacting them.
Read the exact line
Privacy policy
“the right to opt-out of using your Personal Data for training our models or optimizing our technologies.”
For example, you email privacy@deepseek.com and ask them to stop using your chats for training.
They collect voice and photos but say they will not turn them into fingerprints to identify you.
Read the exact line
Privacy policy
“we will not extract or mine voiceprint or facial recognition information or other unique biological patterns or characteristics used to identify a specific individual from the voice inputs or photos you provided to us.”
For example, a voice message is stored as a prompt, but they say they will not build a voice ID of you.
They say they do not sell your data or use it for targeted ads.
Read the exact line
Privacy policy
“NOTE WE DO NOT ENGAGE IN TARGETED ADVERTISING, “SELL” PERSONAL DATA OR USE PERSONAL DATA FOR “PROFILING””
For example, they claim your chat topics will not be used to show you personalized ads.
You can copy or delete chats yourself in the app settings.
Read the exact line
Privacy policy
“you may also copy or delete your chat history via your settings.”
For example, you open settings and delete an old conversation so it is no longer in your history.
On the open platform, you keep your prompts and they give you rights in the model’s answers.
Read the exact line
Terms of service
“You retain any rights, title, and interests—if any—in the Inputs you submit; (2) We assign any rights, title, and interests—if any—in the Outputs of the Services to you”
For example, code the API writes for your app is assigned to you, not kept as their property.
Words to know
Legal words from the lines above, in plain English.
- inputs and outputs
- Inputs are what you type, say or upload to an AI; outputs are what it gives back to you. For example, the question you ask a chatbot is an input and its answer is an output, and both may be stored.
- model training
- Using your content and conversations as examples to teach an AI system, which can then echo them in future answers. For example, a story you write in a chatbot may be studied by the company to make the next version of the bot.
- personal data
- Any information that is about you or can be linked to you, from your name to your phone's ID. For example, your email, your IP address and the list of apps on your phone are all personal data.
- service providers
- Outside companies that handle your data on the app's instructions, such as a cloud host or an email sender. For example, the company that stores the app's files in the cloud sees your data but only does what the app tells it.
- opt out
- Something is on by default and stays on until you find the setting and turn it off. For example, your chats may be used for training unless you go into settings and switch it off.
- indemnify
- You promise to pay the company's legal costs and losses if your use of the service gets it sued. For example, if you post a song you do not own and the label sues the app, you owe the app's lawyer bills.
- affiliates
- Other companies owned by or connected to the same parent company. For example, if you use one app, its sister apps under the same owner may also get your data.
- as is
- You get the service in whatever state it is in, with no promise that it works or will keep working. For example, if the app deletes your photos by mistake, it is not promising to fix that or pay you back.
- warranty
- A promise that a product will work as described; most apps say they make no such promise at all. For example, if a paid feature never works on your phone, a no-warranty clause says that is your problem.
- biometric
- Measurements of your body that identify you, like your face, fingerprint or voice. For example, a selfie used to unlock the app or a voice recording matched to you.
- targeted advertising
- Ads chosen for you based on what you have done across other apps and websites, not just this one. For example, you look at shoes in one app and see shoe ads in a totally different app the next day.
- sell or share
- Under California law, selling means passing your data to others for money or other value; sharing means passing it on for targeted ads. For example, letting an ad network use your browsing history in return for ad space counts as a sale.
How we got here · grade D · score 48/100 · 15 of 15 policy answers backed by a verified quote · 3 not stated
What does the app collect beyond what it needs to work?
They collect account info, your prompts and uploads, device IDs, IP, cookies, usage logs, and approximate location from IP.
Does it record your voice, face or body, and what happens to that?
They collect voice input and photos as content, keep them with other inputs, but say they will not extract voiceprints or face IDs.
Are your chats and uploads used to train AI models, and is that off by default?
Training on your personal data is on by default. You can opt out by emailing them.
Can employees or contractors read your conversations, and when?
Not stated in the documents.
Are you profiled or tracked for advertising?
They say they do not engage in targeted advertising or sell personal data.
Do they sell or share your data, and can you opt out?
They say they do not sell data. They share it with service providers and their corporate group to run and train the service.
What rights do they take over what you type and what the AI makes?
Open-platform terms say you keep rights in inputs and they assign rights in outputs to you. Consumer app license is not in these documents.
How long do they keep your data after you delete it, and can you delete it in the app?
You can delete chat history in settings. Broader deletion is by emailing them. They keep data while you have an account, with no set wipe deadline.
Does it build a lasting memory or profile of you, and can you see, edit or turn it off?
Not stated in the documents.
Do they commit to basic security, and have they leaked data?
They describe commercially reasonable security measures in general terms, with no audit or leak history stated.
Is there a real age gate, and are teens protected?
They say the services are not aimed at children and mention guardians for under-18s, but there is no strong age-check described.
Will they tell you when the rules change, and is the policy specific?
The privacy policy is dated. They may only update the date at the top rather than give advance personal notice.
Can they close your account without warning, and can you get your data out first?
Open-platform terms let them lock or close accounts without prior notice if they think you broke the rules.
If something goes wrong, who pays?
Services are as-is, liability is capped at fees paid in the last year, and you must indemnify them.
Do subscriptions renew on their own, and can you get a refund?
Not stated in the documents.
The privacy policy is 4,248 words at a college level (Flesch reading ease 41.4); the terms are 3,468 words at a college graduate level.
