When you upload a file with ChatGPT, many people may be concerned about how that data will be used for OpenAI model training. Specifically, the policies for personal and business plans are different, so it is important to understand these differences accurately.
In this article, we will explain the possibility that files uploaded based on the official help will be used for model training and our efforts to protect privacy.
table of contents
- table of contents
- ChatGPT is learning a huge amount of data
- Personal data usage
- Data usage for business
- How uploaded files are used
- Cases of use in learning
- If you are excluded from learning
- What you can do to protect your privacy
- Be careful when handling authoritative information
- Consider opting out of learning in your account settings
- Consider using the business version/API
- summary
ChatGPT is learning a huge amount of data
Contains ChatGPTLarge-scale language models (LLMs) are mechanisms for natural language understanding and sentence generation by incorporating huge amounts of text data.is. Publicly available information on the Internet and various licensed datasets are used for training, resulting in the ability to cover a wide variety of topics.
However, the data used in these learning processes can also include user-uploaded files and text input. What data is used for learning depends on the type of service provided, contract plan, opt-out settings, etc.
For this reason,It is important to manage confidential and personal information so that it is not used for model learning and understand usage policies.It will be. From here, let's take a look at how files uploaded to ChatGPT are handled.
Personal data usage
If you mainly use personal services such as ChatGPT or DALL·E,The content you upload may be used to improve the modelThere is.
Examples of applicable usage scenarios
If you are using ChatGPT free plan or ChatGPT Plus for personal use
If you upload files for personal use within your account
OpenAIの[公式ヘルプ](<https://help.openai.com/en/>)では、「個人向けに提供しているサービスでは、ファイルを含むアップロードされたコンテンツをモデル学習に活用する可能性がある」と明言されています。
However, OpenAI provides a mechanism for users to manage their own settings through the data control function. If necessary, consider opting out of ChatGPT settings to "not use it for model training."
Data usage for business
on the other hand,Business useWhen using ChatGPT Enterprise or OpenAI API assumingUploaded files or text are not used to train the model.
Examples of applicable usage scenarios
If you have a corporate contract with ChatGPT Enterprise
If you are using OpenAI API to link with your company's system and process files and text independently.
According to official information, OpenAI's policy is that no data exchanged through enterprise plans or API usage will be used for model training.
This reduces the risk of confidential company information and users' personal information being diverted to other parties for model improvement.
How uploaded files are used
How uploaded data is processed depends on the type of service provided by OpenAI and how it is used.
Cases of use in learning
In personal services, usersData uploaded to ChatGPT and entered text are saved on OpenAI's server and then used to improve the model according to certain rules and periods.There are cases where this happens. However, this is done automatically, using the following techniques:
Anonymization/aggregation
Remove as much personal information and identifiable data as possible from users and incorporate overall trends into model learning
Compliance with Terms of Use and Privacy Policy
Transparent how data is used in accordance with OpenAI's privacy policy
If you are excluded from learning
If you are using a business service, data learning as described above will not occur.. This is a major security advantage when companies handle authoritative data.
What you can do to protect your privacy
Understanding how personal and confidential information is handled when using ChatGPT is the first step to using the service with peace of mind. Here are some concrete steps you can take to protect your data.
Be careful when handling authoritative information
parableEven when using ChatGPT for individuals, avoid unintentionally uploading files containing highly confidential or personal information.. In particular, it is important to carefully examine the contents of contracts, ID information, and the original text of research data before uploading them.
Consider opting out of learning in your account settings
The ChatGPT settings screen provides choices regarding data usage with ChatGPT. If your information is sensitive, you can limit its use for learning by opting out of the use of your data. You can toggle "On" or "Off" from "Improve models for everyone" in the "Data Control" tab of the settings screen.
Consider using the business version/API
If your company or organization requires a high level of privacy protection, one option is to consider introducing ChatGPT Enterprise or OpenAI API. No data is used for model training, including file uploads, so you can operate with confidence.
summary
Whether the files and texts you upload with ChatGPT are used for learning depends largely on whether you have a personal or business plan. In ChatGPT for individuals, files uploaded by users may be used to improve models, but in Enterprise plans and APIs for companies, data is not used for learning.
If you want to protect your privacy and confidential information, consider options such as not uploading files containing sensitive information, opting out of data use for learning from the settings screen, or using business services. By knowing these policies, you should be able to use ChatGPT with more peace of mind.