Privacy policy for the use of ChatGPT: A guide
ChatGPT has been keeping companies, politicians, the media and, of course, users busy since November 2022. In addition to the many new possibilities and opportunities that the ChatGPT chatbot brings with it, the issue of data protection and data security naturally also plays an important role. In this article you will find the most important information and facts about data protection when using ChatGPT and the ChatGPT API.
What is ChatGPT?
ChatGPT is a very advanced chatbot developed by the US company OpenAI. The company makes a free version of the chatbot available to all interested users via the website or app. Users can also purchase a subscription for the Pro version to benefit from additional features. ChatGPT simulates human-like conversations, generates texts and interacts with its users in real time. The chatbot can interact with people in real time, answer questions, write texts or hold general conversations. The special thing about ChatGPT is above all the way it conducts human-like dialogs and writes texts almost independently. The chatbot can even handle context.
For all those who have never tested ChatGPT themselves, here is a short video of ChatGPT writing a birthday invitation.
You can read more about how to use ChatGPT and how to give it the right commands in my article on ChatGPT prompts.
How does ChatGPT work?
ChatGPT is further developed through an iterative process of machine learning and continuous improvement. It is based on training with large amounts of text data generated by humans. The model learns by analyzing this data and recognizing patterns, correlations and linguistic structures. This analysis trains the model to understand texts and generate suitable answers. OpenAI also uses a method called Reinforcement Learning from Human Feedback (RLHF) to improve the model. Human experts are used to evaluate the generated answers and provide feedback. Based on this feedback, the model is fine-tuned and optimized to provide more precise and higher quality answers. This continuous learning and improvement process allows ChatGPT to expand its capabilities, absorb new information and current developments, and provide an ever more natural and accurate conversation. OpenAI is actively working to further develop ChatGPT and continuously improve its performance to meet the needs of users.
The privacy policy of ChatGPT
What information is collected and how is it used?
When asked whether OpenAI stores the data that users type into the chatbot, the company gives the following answer:
The data transmitted via the OpenAI API is not used to train OpenAI models or to improve the OpenAI service offering. The data transmitted via the non-API consumer services ChatGPT or DALL-E can be used to improve our models. You can find all further information on this page: https://openai.com/security
In other words, OpenAI stores personal data to improve the chatbot ChatGPT and train the language model in the background. OpenAI even admits that they do not rule out sharing the data with potential partners if this serves the further development of the chatbot.
I therefore recommend that you do not type any private or business data into the chatbot.
Measures to protect the privacy of users
There are currently two ways in which you as a ChatGPT user can prevent your data from being stored or used:
- You can opt out of having your data used to improve OpenAI’s non-API services by filling out this form with your organization ID and the email address associated with the account holder.
- OpenAI states that it will delete your content (such as prompts, generated images, uploads and API responses) if you submit a data deletion request. According to OpenAI, processing a request for data erasure can take up to 30 days after receipt of the request.
Apart from the fact that you actively prevent OpenAI from saving your data, you can of course simply refrain from entering personal data when using it. Always make your requests anonymously and avoid typing personal data such as addresses, account information or similar into the chatbot.
Data storage and deletion at ChatGPT
How are users’ chat logs and personal data stored?
OpenAI initially saves all conversations that users have with the chatbot ChatGPT. The company also admits that this data is used for further development of the chatbot and the platform. If users want to chat with ChatGPT, they must first agree that they understand that OpenAI saves the chat entries (see image).
Guidelines on the deletion of data and retention period
As a ChatGPT user, you have the right to have your data or entries deleted. If you wish, you can submit a request for data deletion to OpenAI. OpenAI then has 30 days to delete your data.
Data protection with the ChatGPT API
What is the difference between ChatGPT and the ChatGPT API?
OpenAI distinguishes between ChatGPT and the ChatGPT API in its privacy policy. So what’s the difference? ChatGPT refers to the actual model developed by OpenAI. It is based on the GPT 3.5 architecture and is an advanced AI model that is able to simulate human-like conversations and generate natural language texts. ChatGPT can be used for various purposes, including customer support, virtual assistants and general conversations.
The ChatGPT API (Application Programming Interface), on the other hand, is an interface provided by OpenAI. It enables developers to access the ChatGPT model and use its functions in their applications. The API acts as an intermediary between the application and the ChatGPT model so that requests can be sent to the model and the generated responses are returned. The ChatGPT API allows developers to utilize the power of ChatGPT in their own products and services without having to implement or train the model themselves.
To summarize, ChatGPT is the actual AI model, while the ChatGPT API is the interface that enables access to the model and facilitates interaction with the AI.
If you use the well-known chatbot ChatGPT via your web browser or via the smartphone app, you use ChatGPT and OpenAI saves your data. However, when companies offer their own services based on the ChatGPT API, it is precisely the ChatGPT API that is used, for which OpenAI states that it does not store any data.
What do companies need to consider when using the ChatGPT API?
If users not only want to use the public chatbot ChatGPT, but also want to offer their own services based on the ChatGPT language model, they can use the ChatGPT API. Companies such as Helvetia Switzerland, for example, have expanded their own chatbots with the help of the ChatGPT API. In this scenario, the responsibility for data protection and informing users about the privacy policy lies with the company using the ChatGPT API. The following example from Helvetia Switzerland’s ChatGPT experiment illustrates this.
ChatGPT and copyright
When does copyright apply?
Copyright and intellectual property issues may arise if content created by users is shared outside the platform without the permission of the author or creator. ChatGPT is able to write texts on various topics. It was trained by OpenAI by analyzing a large amount of text documents. These documents come from various sources.
Plagiarism occurs when third-party intellectual property is used without consent or reference to the source. Although ChatGPT as an artificial intelligence has no intellectual property of its own, the responses generated by ChatGPT could still be considered plagiarized if the underlying training data contains plagiarized work. So, when using ChatGPT, continue to consider source references and if you are unsure, research which sources ChatGPT bases its answers on to be sure.
When is ChatGPT intellectual property?
The issue of copyright opens up a further dimension. How is the content generated by ChatGPT handled? Is there a possibility that they will be considered valuable works under copyright law in the future?
In principle, copyright law only protects works that are based on human creative creation. However, since ChatGPT is an artificial intelligence and not a human being, it lacks human creative creation when humans simply ask ChatGPT to create a simple text. However, if ChatGPT’sprompt engineering response was generated by a human based on a complex and creative prompt, it could be argued that this is a human creative creation.
Is there a law against ChatGPT?
Countries such as Italy have banned the use of ChatGPT. In Switzerland, organizations such as Swiss Fintech Innovations have been looking into the legal situation regarding ChatGPT and artificial intelligence. According to Cornelia Stengel, Co-Director of Swiss Fintech Innovations, there is still no law that prohibits the use of language models such as ChatGPT. Instead, companies must take other individual and specific guidelines into account (see screenshot).
Tips for protecting your privacy when using ChatGPT
- If you are a user of the chatbot ChatGPT, please avoid entering personal information.
- If you are unsure whether you have given the chatbot ChatGPT too much personal data, take the opportunity to delete your data and request OpenAI to delete your data immediately,
- If you as a company offer services based on the ChatGPT API, you are responsible for the data protection of your users. Ensures that no data is disclosed to third parties.
- When using the ChatGPT API, don’t forget to inform users about the opportunities and risks of this API service right from the start.
Conclusion: Why data protection should not be underestimated when using ChatGPT
ChatGPT has already developed into an extremely powerful and widely used tool. The number of users clearly exceeds that of other digital platforms. Given the wide range of functions, it is not surprising that more and more users and companies are using the chatbot. The use of ChatGPT is by no means prohibited. On the contrary, I believe that we should definitely make use of the possibilities of this new chatbot and benefit from it. However, we should not completely ignore the issue of data protection. As with every platform, there are certain “rules of the game” with regard to data protection and the handling of sensitive data. As long as these are taken into account, there is nothing to prevent the responsible use of ChatGPT.
Privacy policy for the use of ChatGPT – Frequently asked questions & answers
What exactly is ChatGPT?
ChatGPT is an advanced AI-driven text generator from OpenAI that can simulate human-like conversations.
How does ChatGPT collect my data?
ChatGPT collects the texts you enter during a session in order to process and respond to the interactions.
Does ChatGPT save my personal conversations?
Since the release of GPT-3, OpenAI does not store the calls permanently, they are only kept for a limited period of time and then deleted.
Can ChatGPT share my data with third parties?
OpenAI’s privacy policy prohibits the disclosure of your personal information to third parties unless required by law or necessary for the provision of the service.
What happens to my data when I stop using ChatGPT?
OpenAI has data retention and deletion policies that govern how data is handled after use has ended.
What rights do I have with regard to my data at ChatGPT?
You have the right to view, correct, delete and restrict the processing of your data. You also have the right to data portability and the right to lodge a complaint with a supervisory authority.
How will ChatGPT keep my data secure?
OpenAI implements a variety of security measures, including physical, electronic and administrative security measures, to protect your data and ensure its confidentiality, integrity and availability.
How can I protect my privacy when using ChatGPT?
Avoid entering sensitive personal information, use pseudonyms where possible and be aware that anything you enter will be treated as data that ChatGPT uses to interact with you.
