In short
- Uploading blood test results to ChatGPT is much safer once you remove everything that identifies you and switch off model training. The values alone say little about who you are.
- To stop training, turn off "Improve the model for everyone" in Settings, Data Controls. Or use Temporary Chat, which is not used for training while it stays temporary.
- A deleted chat is scheduled for permanent deletion within 30 days, with security and legal exceptions. Files saved to Library are not deleted with the chat.
- According to OpenAI's help page, Health in ChatGPT is available to eligible users in the U.S. It does not list the EU or Slovenia.
- Under the EU's GDPR, data concerning health is a special category. No chatbot answer is a diagnosis.
Your report is one of the most personal documents you have. This guide explains what happens to a file you give ChatGPT, which settings to change, and exactly what to black out first. To write a good question afterwards, use our ChatGPT prompt template for blood test results.
Is it safe to upload blood test results to ChatGPT?
It can be reasonably safe if you remove your identity first, turn off model training or use Temporary Chat, and delete the chat and file when you are done. The risk sits mostly in your name, birth date and ID numbers, not in the values.
A ferritin of 22 µg/L or an LDL of 3.4 mmol/L does not reveal who you are. Your name, date of birth, insurance number and a barcode do. So the safest upload is one where the report no longer points to you.
Safety also means the answer. A chatbot can misread a scan, guess a range or change its answer. In a 2025 study, the lower reference limits ChatGPT gave for the same test varied with a median of 26.5 % across repeats. Our article on what an AI blood test analyzer can and cannot do explains more.
What happens to a file you upload to ChatGPT?
Your chat is saved to your account until you delete it, and uploaded files are saved to your Library when Library is available for your account. OpenAI's help page says chats and Library files are managed separately.
That detail matters. According to OpenAI, deleting a chat does not delete files saved to Library. You view and delete saved files in Library itself.
When you delete a chat, it is removed from your account at once. It is then scheduled for permanent deletion from OpenAI's systems within 30 days. OpenAI names two exceptions: the chat was already de-identified, or it must be kept longer for security or legal reasons.
Files you upload to a project or a custom GPT are kept until you delete that project or GPT. After deletion, they are removed within 30 days, with the same legal and security exceptions.
How do you stop ChatGPT from training on your blood test?
Turn off "Improve the model for everyone" in your data settings. OpenAI states that your chats then stay in your history but are not used to train its models.
- On the web: click your profile icon, open Settings, go to Data Controls and turn off "Improve the model for everyone".
- On mobile: open the side menu, tap your profile icon, choose Data Controls and turn off the same setting.
- Signed out: click the question mark icon at the bottom right, open Settings and turn it off.
The setting applies to your whole account, on every device, and you can change it at any time. Labels may appear translated if your app is not in English.
What does Temporary Chat change?
Temporary Chat is a one-off conversation. Start it with the "Temporary" button at the top right of a new chat. OpenAI says it does not appear in your history and does not create new memories. It is not used for training while it stays temporary.
Two limits apply. OpenAI may keep a copy for up to 30 days for safety purposes. A personalised Temporary Chat can also use your existing memories, and if you save the chat, it becomes a regular chat.
Is ChatGPT Health available in the EU and Slovenia?
Not according to OpenAI's help page as of September 2026. It states that Health "is available to eligible, logged-in ChatGPT Free, Go, Plus, and Pro users in the U.S. who are age 18 or older".
When OpenAI announced ChatGPT Health in January 2026, early access was open outside the European Economic Area, Switzerland and the United Kingdom. A July 2026 update on the same page says Health is launching to U.S. users aged 18 and older. Availability can change, so check your own app.
Health is a separate space. OpenAI says connected medical records and Apple Health data, and conversations that use them, are not used to train its foundation models. A normal chat without connected Health data follows your usual training setting. OpenAI also states that Health is not intended for diagnosis or treatment.
What should you remove before you upload a blood test?
Remove everything that points to you, and keep only what the explanation needs: age, sex, values, units and your lab's ranges. The table shows what to do with each part of a typical report.
| Part of the report | Remove or keep | Why |
|---|---|---|
| Full name | Remove | It identifies you directly and adds nothing to the explanation. |
| Date of birth | Remove, give age in years | Some ranges depend on age, but the exact birth date is not needed. |
| National ID and insurance number (in Slovenia, EMŠO and the ZZZS card number) | Remove | These numbers are unique to you. |
| Address, phone and email | Remove | They identify you and are not medical information. |
| Barcode, QR code, sample and request numbers | Remove | They can link the report back to your record at the lab. |
| Doctor's name and ID, lab name and ID numbers | Remove | Together with a date, they narrow down who you are. |
| Sex | Keep | WHO sets different anemia cutoffs for hemoglobin in women and men. |
| Values, units and the lab's reference ranges | Keep | The explanation depends on them, for example ferritin in µg/L or TSH in mIU/L. |
Source: OpenAI Help Center, Chat and File Retention Policies in ChatGPT (2026); hemoglobin cutoffs from WHO 2024. Reference ranges differ between laboratories. This is practical guidance, not legal advice.
Screenshot, PDF or typed text?
Typed values are the cleanest option, because you share only what you choose. A screenshot of a page where you blacked out personal fields is the next best.
Be careful with a PDF. Some PDF editors draw a black box over text but leave the text in the file. A PDF's file name and properties can also contain your name. Photos from a phone can carry extra data, such as the location. If in doubt, take a screenshot of the covered page, or type the values.
Values such as ferritin, TSH, hemoglobin and LDL cholesterol are fine to share without your identity. For what they mean, see how to read a blood test report.
What does EU law say about health data?
The EU's General Data Protection Regulation (GDPR) treats data concerning health as a special category. Article 9 prohibits processing it unless an exception applies, and explicit consent is one of those exceptions.
Article 4(15) defines data concerning health as personal data about a person's physical or mental health that reveal information about their health status. A blood test with your name on it clearly fits.
These rules bind the companies that process your data. They do not stop you from sharing your own report. That choice is yours, which is why removing your identity is the most practical protection you control.
A chatbot is also not your doctor. OpenAI states that Health is not intended for diagnosis or treatment. Some causes of symptoms never show in blood at all; see what a blood test cannot tell you.
A private chat is still not medical care. Never start, stop or change a medicine or dose because of a chatbot answer. If a result comes with chest pain, breathlessness, fainting, confusion, new weakness, black stools or heavy bleeding, get medical help now. In an emergency, call 112 or your local emergency number.
How Blood Lab 360 reads this
Blood Lab 360 parses and rates your report on your device. The ratings, the plan and the trend are stored in your browser, not on our servers. It needs only the numbers, your age and your sex, and you can black out your name and birth date before uploading.
Blood Lab 360 is run by Feel Awsm d.o.o., an EU company in Ljubljana, Slovenia. Your explicit consent under Art. 9(2)(a) GDPR is recorded with a timestamp and version. You can withdraw consent at any time in the app, and everything is then deleted, including reports, plan and chat history.
If you use the AI chat, your measured values are sent to an AI processor to compose the answer. You give separate consent for this and can withdraw it. The app does not diagnose, does not sell your data, and does not guess a range when your age or sex is missing.
Frequently asked questions
Can I upload my blood test results to ChatGPT?
Yes, ChatGPT accepts documents and images. Before you upload, remove your name, birth date, ID and insurance numbers, address and barcodes. Consider turning off model training or using Temporary Chat.
Is it safe to upload a blood test to ChatGPT free?
The same privacy steps apply on the free plan. OpenAI's Data Controls let you turn off "Improve the model for everyone", even when signed out. Remove your identity first either way.
Is it safe to upload blood test results to AI?
With any AI tool, check three things first: whether your chats train the model, how long files are kept, and how to delete them. Remove your identity before uploading, whichever tool you use.
What is ChatGPT Health for blood test results?
Health is a separate space in ChatGPT that can use connected records to explain recent labs. OpenAI's help page lists it for eligible users in the U.S. aged 18 or older. It is not intended for diagnosis or treatment.
Does deleting the chat delete my uploaded file?
Not always. OpenAI says deleting a chat does not delete files saved to Library, so delete those in Library too. Deleted chats are scheduled for permanent deletion within 30 days, with legal and security exceptions.
