Cybersecurity 
Modern law firms handle vast amounts of confidential client data, including Personally Identifiable Information (PII), financial records, case files, sensitive business documents, court filings and more.
7 min read
Published on July 23, 2024
Updated on May 29, 2025.
Depending on how you use it, ChatGPT can be safe. ChatGPT is a Large Language Model (LLM) developed by OpenAI and trained to produce human-like text. Based on GPT-3.5, ChatGPT is an advanced Artificial Intelligence (AI) model whose database is formed from massive amounts of text data. People often use ChatGPT to simplify complex topics, generate ideas, write code and develop marketing material. However, to protect yourself and your data, you should be aware of some security risks, including privacy concerns, copycat ChatGPT websites and the tool’s tendency to generate misinformation.
Continue reading to learn what risks users face when using ChatGPT and how to stay safe while using it.
ChatGPT safety measures
Here are some security measures that ChatGPT follows to protect your privacy:
- Content moderation: OpenAI uses a combination of advanced algorithms and human reviews to help filter out hateful, offensive and inappropriate content. This helps prevent the spread of misinformation, violent content and harmful outputs.
- Security audits: ChatGPT undergoes an annual security audit to identify potential vulnerabilities. These regular reviews are made to identify and address security vulnerabilities in the system’s software before they can be exploited.
- Strict access controls: Only authorized users can access the inner workings of ChatGPT through several measures, including Role-Based Access Controls (RBAC) and Multi-Factor Authentication (MFA) methods.
- Data encryption: OpenAI uses strict encryption protocols to protect all data between a user and ChatGPT from unauthorized access. This helps protect user inputs and outputs from being intercepted by third parties.
- Bug bounty program: ChatGPT has a bug bounty program, which encourages ethical hackers and security researchers to test the system and report any bugs they find. Verified findings are rewarded, allowing OpenAI to improve its security through the community’s help.
ChatGPT security risks
Although OpenAI has various safety measures in place, some potential risks associated with using the generative AI solution could jeopardize your personal information.
Privacy concerns
One of the main security concerns when using ChatGPT is the potential for privacy violations. According to OpenAI’s privacy policy, under specific circumstances, ChatGPT may share user data with third parties without notifying the user. The information you input when using ChatGPT is saved, stored and potentially shared with external sources. In addition to saving the content of your conversations, ChatGPT also stores data associated with you and your device. When reading a privacy policy, it’s important to understand how ChatGPT collects, stores and uses your data. Since ChatGPT isn’t confidential unless you turn off your chat history, all conversations you have with the chatbot are stored to help improve the AI model.
Illegitimate ChatGPT websites and apps
As ChatGPT gained popularity, scammers started creating fake websites impersonating ChatGPT as a way to steal personal information. In April 2023, Google Play found hundreds of illegitimate ChatGPT apps, many of which were downloaded by millions of people, exposing their private data to cybercriminals. Always use ChatGPT’s official website and app to protect your information from being unintentionally shared with cybercriminals through fraudulent websites and apps.
Misinformation
Although ChatGPT can produce human-like content, its responses aren’t always accurate. ChatGPT has been known to demonstrate a chatbot behavior called hallucination, in which it spontaneously generates information that appears true but is actually false. Because of this phenomenon, fully relying on ChatGPT for accurate information can be risky. In addition to occasionally generating false information, ChatGPT may also reflect biases and stereotypes present in the data on which it was trained. For these reasons, don’t use ChatGPT as a primary source for research, but instead as a starting point for guidance. Any content provided by ChatGPT should be thoroughly reviewed and fact-checked by a knowledgeable expert or credible source in the subject area.
Malicious use
Another important security concern with ChatGPT is its potential to be used maliciously. While ChatGPT is designed to support productivity, the tool can also be exploited by cybercriminals for malicious purposes. For example, ChatGPT can be used to write convincing phishing emails or assist in writing malicious code with alarming accuracy. Luckily, OpenAI has reduced these risks through content filters and usage policies, but cybercriminals can still find ways to bypass security measures.
Tips to stay safe when using ChatGPT
Despite some privacy concerns and security risks, there are ways to use ChatGPT safely. Follow the tips below to stay safe when using the tool.
Only use ChatGPT’s official website and app
Because ChatGPT is so popular, cybercriminals tend to create fake websites and apps that impersonate the real platform in an attempt to steal your information. To stay safe, make sure to use only ChatGPT’s official website and verified app. Avoid downloading third-party versions of ChatGPT or clicking on suspicious links that resemble ChatGPT’s, especially if they are shared through email or social media. If you’re worried about accidentally visiting an impostor website, bookmark ChatGPT’s official website so it is easier to access. This ensures you always reach the legitimate platform and helps protect your data from phishing attempts.
Never enter sensitive information into ChatGPT
ChatGPT’s privacy policy acknowledges that there are certain occasions when your private information may be shared with third parties. Additionally, all input provided to ChatGPT is stored, which means your interactions aren’t completely private. If ChatGPT suffers a data breach, any sensitive information you’ve shared could end up in the wrong hands. This is why it is crucial not to upload sensitive documents like legal PDFs or financial records into ChatGPT.
Don’t use ChatGPT to create passwords
You should not use ChatGPT to generate passwords for several reasons. First, any content produced by ChatGPT stays in OpenAI’s database. If it creates passwords for you and ChatGPT is breached, a cybercriminal could gain access to those stored passwords. Second, since ChatGPT is designed to generate text based on patterns instead of complete randomness, it might produce similar or even identical passwords for multiple users.
A safer and more reliable method is to use a password generator and password manager. A password generator creates strong, unique passwords using a random combination of uppercase and lowercase letters, numbers and symbols. Once you’ve generated a secure password, a password manager can safely store it for you. Tools like Keeper Password Manager feature a built-in password generator so that you can create and store your passwords in the same place.
Use an anonymous account
If you want to minimize security risks when using ChatGPT, it is safest to use an anonymous account. In 2024, ChatGPT started allowing people to use its chatbot instantly without creating an account. This option helps protect your privacy by removing the need to share personal information, like your full name and email address. Even if you choose to create an account, you can limit data sharing by turning off the Improve the model for everyone option in ChatGPT’s settings. Turning this off will prevent your interactions from being used to train the AI model. However, you may still be required to provide your phone number for authentication purposes.
Always cross-check the information ChatGPT provides
Since ChatGPT can sometimes provide inaccurate information through hallucinations, always cross-check its responses with reliable sources. Verifying facts through additional research not only helps confirm ChatGPT’s accuracy but also allows you to identify potential biases in how the information is presented. Using ChatGPT as a base for research is fine, but independent research is crucial to ensuring the information you use is accurate and unbiased.
Report issues you encounter to OpenAI
When using ChatGPT, you may come across some issues that need to be resolved, such as incorrect responses or unexpected behavior. Contact OpenAI directly so the team can identify and fix bugs quickly. After receiving ChatGPT’s response to your prompt, you can also give immediate feedback based on how satisfied you are with the content. If the response is inaccurate or unhelpful, you can give it a thumbs down to flag the issue. In contrast, if you find ChatGPT to be convenient and easy to use, you can give the response a thumbs up. This feedback helps OpenAI improve its model over time, ensuring a better experience for all users.
Use ChatGPT and other AI models safely
Whether you’re using ChatGPT or other AI tools, it’s important to prioritize safety and privacy. Make sure you use the official ChatGPT website or app, and double-check the accuracy of ChatGPT’s results using trusted sources. As AI continues to learn from user interactions, stay informed and cautious to ensure you benefit from the technology. By following best practices, you can use AI tools like ChatGPT responsibly and safely.
Frequently asked questions
Is ChatGPT encrypted?
Yes, ChatGPT uses encryption to protect data in transit. All communication between you and ChatGPT is secured using HTTPS, which encrypts data to prevent unauthorized access while it’s being transmitted. This means the information you input and receive is protected from being intercepted by third parties during transmission. However, even though the connection is encrypted, the data you input into ChatGPT may be stored and used unless you disable certain settings designed to improve the model.
Should I trust ChatGPT?
You can trust ChatGPT as a tool for generating ideas or explaining challenging concepts, but it’s important to use it with caution. ChatGPT can sometimes generate incorrect responses through a phenomenon called a hallucination, and it may reflect biases found in its training data. Although ChatGPT can be a valuable tool, it shouldn’t replace human judgment, primary research sources or expert advice.
Can I delete my ChatGPT history?
Yes, you can delete your ChatGPT history. OpenAI allows users to manage and delete their past conversations from their account settings. You can remove chats or clear your entire history by clicking on your profile icon, going to Settings, selecting Data Controls, turning off chat history and deleting existing conversations. By disabling chat history, you prevent your future conversations from being saved or used to improve the AI model. However, some data may still be saved for security monitoring purposes, as defined by OpenAI’s privacy policy.
