After the ChatGPT’s release, its parent organization OpenAI has welcomed grave objections from privacy specialists & users due to concern over its Data Privacy policies. This article will uncover some major ChatGPT Data privacy concerns to make you understand the significance of taking additional safety measures while using this AI chatbot.
What are Chatgpt Data Privacy Concerns?
Let’s understand different ChatGPT privacy concerns, which puts its overall usage into question. Using ChatGPT brings risks to users, which are as follows-
User privacy
Chat GPT-3 and 4 architecture models were trained on an enormous amount of people’s data from the internet. Experts are still wondering how OpenAI collected this data. This gathered data includes people’s personal information like addresses, names, locations, etc., whose permission was not taken from them.
As stated by OpenAI in their privacy policies, they can share users’ Personal Information with their business partners, vendors, service providers, affiliates, third parties, or even before the law without taking permission from users, who can face severe consequences if their data gets misused.
Further, ChatGPT stores your conversations or every text or image you share. OpenAI supports this action in its privacy policy by using this collected info for training the AI bot with new data. What if you enter any confidential information about your work? This way, your personal data is getting compromised for the profit of OpenAI, putting it at cybersecurity and data breach risks.
Bias and discrimination
As we know, the ChatGPT AI bot offers answers based on data and information it was trained with. Many users have discovered that the bot can give biased answers that discriminate against races, genders, minority groups, etc.
The unfair practices and biased perspectives can be reflected in the bot’s answers, which OpenAI must have considered while developing ChatGPT. OpenAI has said that they will fix this issue as they will gather more feedback from users who can flag the bot’s biased results.
Misuse of data
We have described all types of personal data OpenAI collects from users. Most data includes standard information that almost every website collects. The significant risk is to your information shared via conversations. Even if you delete your conversations history, OpenAI may still keep them on its servers. Doesn’t that sound alarming?
Your personally collected information is visible to many people without sending any notice – OpenAI employees, vendors, service providers, business partners, affiliates, law enforcement authorities, or other third parties for different purposes.
OpenAI strictly asks its partners to follow its privacy laws and keep the overall right under its control. There’s never a 100% guarantee that your data cannot be misused at any time.
The March 2023 Data breach incident in the software bug of ChatGPT put the payment-related information of 1.2% of plus-service subscribers into open risk for misuse, leaving a thousand others vulnerable. Although OpenAI addressed and disclosed all related information about this incident. We can’t deny that sensitive data got disclosed to many other users.
Consent
Using personal information online without consent from people has put OpenAI at risk. Italy’s data regulator in March administered a quick emergency decision instructing OpenAI to stop utilizing millions of Italians’ personal details as training data. Many global artists & media organizations have groaned about OpenAI using their original copyrighted works without authorization.
OpenAI’s Data privacy policy doesn’t explain its legal explanations for utilizing people’s personal details in training data. Still, it supports it by saying it depends upon “fair interests” when creating its services.
ChatGPT might disobey international regulations such as GDPR and CCPA data security laws when collecting information without consent. Even if people’s private information is publicly available, their unauthorized use can break the fundamental principle of contextual integrity used for legal privacy arguments.
Data Ownership
Data ownership is still a debated topic for ChatGPT. The scratched ChatGPT training data can be copyrighted or proprietary. The ChatGPT AI tool produces copyrighted answers from different sources, as tested by many users. Many users, including website owners, companies, artists, or other individuals, are not paid or compensated by OpenAI.
Data retention
ChatGPT registers every input instruction or message every user provides. Per the “Security & Retention” section of the OpenAI Privacy policy, they can retain your data for different purposes, such as business, research, resolving disputes, security reasons, and to comply with legal responsibilities indefinitely as they want without sending the users any notice.
Even if you delete your conversation threads from your ChatGPT account, it only
goes away from OpenAI servers once they want. OpenAI allows users to Opt-out by sharing their personal information with them by filling out forms. We can’t neglect the confidential information of users getting retained.
International regulations
As per Europe’s GDPR (General Data Protection Regulation) laws, Anyone cannot use people’s personal data available online, even if it’s free without permission, which can be any information from their name to an IP address or needed to recognize a person. According to Italy’s data regulator Garante, OpenAI’s ChatGPT doesn’t comply with GDPR rules because of the following reasons-
- OpenAI doesn’t include any measures to stop children under 13 years of age from utilizing its services.
- Much information on some people has been found to need to be more accurate.
- People’s consent wasn’t taken before using their personal data for training.
- OpenAI has not mentioned any legal arguments to support their act of gathering people’s data for ChatGPT training. For a business to organize and utilize people’s data under GDPR, it must depend on one of six legal explanations, from taking consent to mentioning the information needed.
Does ChatGPT collect personal data?
ChatGPT collects two types of Personal Information of users. OpenAI majorly does it to improve its services and train the AI bot to enhance its answers providing accuracy.
1. The Personal Information you provide to ChatGPT after creating an account and while communicating. This includes the following.
- Your Account Information, including name, account credentials, contact details, payment card data, and transaction history.
- The content you share, such as Inputs, uploaded files, and feedback while using ChatGPT services.
- Information you provide to ChatGPT when you communicate with this platform.
- Your Social media information, like contact details and more, when interacting with ChatGPT’s official social media pages on Twitter, Instagram, Facebook, Youtube, LinkedIn, and Medium.
2. Personal Information when you use and interact with ChatGPT services. This includes-
- OpenAI receives your Log data (browser type, IP address, request’s time & date ) through browsers automatically whenever you use ChatGPT’s website.
- Your Usage Data automatically reaches OpenAI, which consists of information about the type of content you see, engage with, and features used, and your country name, device type, time zone, etc.
- Your Device Information, including its name, browser used, and OS.
- Your device’s cookies to personalize your experience.
How is Chatgpt a data privacy nightmare?
As per articles and blogs published by OpenAI, They have fed an enormously large amount of data that is 570GB and approx 300 billion words from different GPT-3, 4, and DALL-E-2 data sources, including website pages, articles, research articles, books, Wikipedia, etc.
This was done to make ChatGPT AI bot expert in analyzing patterns and generating human-like responses in seconds. Did OpenAI ever get permission from people worldwide before using their personal data to train ChatGPT AI Bot? Isn’t it a significant concern and a breach of copyright issues?
OpenAI still needs to clarify the exact data used to train the ChatGPT AI bot. The biggest nightmare of users and experts worldwide is how OpenAI uses our personal information fed as input prompts in the conversations with ChatGPT. As mentioned in the Privacy Policy of OpenAI, the company can record our conversation info and data.
Their employees can view your conversations information to train the AI Bot and provide more personalized answers. The next section will give you more details on CharGPT Data Privacy concerns.
