Is ChatGPT Safe to Use or Not? Secure Your Data in 2024

ChatGPT is rapidly changing how we live and work, but it comes with challenges. There are ethical, privacy, and security implications to be wary of as a user. The bot’s parent company, OpenAI, has strict protection policies, but these only help to minimize risks and aren’t foolproof.

Whether or not ChatGPT is safe to use mostly depends on how well you understand the underlying issues with the bot — and the precautions you can take to keep yourself safe while chatting. That’s why I’ve put together a complete guide to the chatbot’s safety and how to secure your interactions. Read along to discover how safe ChatGPT is for personal or work-related use, as well as its implications for young children.

ChatGPT’s Safety Measures

OpenAI has a system of safety rules, controls, and data handling policies to make ChatGPT secure and to help boost user trust. Here are the main protections:

Data Encryption

ChatGPT provides end-to-end encryption for all conversations. This encryption converts data into unintelligible characters that can only be deciphered by the intended recipient. It provides security against unauthorized access or interception of your information while using the chatbot.

Strict access controls are a fundamental security measure used by OpenAI to ensure that only authorized personnel can access sensitive areas of ChatGPT’s infrastructure and code base. This includes controls over who can view or modify the AI model’s code and training data. These measures help protect the integrity and confidentiality of the system, which is crucial for maintaining user trust and compliance with data protection regulations.

Regular System Security Audits

The chatbot undergoes annual security audits by independent cybersecurity specialists to identify and address potential vulnerabilities in the system. This minimizes weak spots that could be exploited to launch cyber attacks or compromise the platform’s users. OpenAI also runs a Bug Bounty Program that invites ethical hackers, security researchers, and tech enthusiasts to identify and report any possible weaknesses.

Data Storage and Retention Policies

Although ChatGPT may collect and store chat data for model training, the platform commits to strict retention policies, including data anonymization. That makes it harder to trace information like chat content to a specific user. The bot also complies with regional data protection regulations — like GDPR — to handle and store your data securely.

For example, Italy initially suspended the use of ChatGPT due to concerns about data privacy and compliance with the European Union’s GDPR. However, ban was reversed after OpenAI addressed those concerns, ensuring that their data processing met regulatory standards.

ChatGPT, like many SaaS applications, collects two types of personal information as outlined in OpenAI's privacy policy. The service automatically gathers device data, usage details like location and time, and log data including IP address and browser type. Additionally, it stores data you provide directly, such as account information and the content of your interactions. This comprehensive data collection helps OpenAI analyze user engagement and enhance the functionality of its AI models, ensuring a more tailored and efficient user experience.

Privacy Settings

OpenAI gives you the ability to manage your interactions with ChatGPT. You can delete chat history and stop ChatGPT from saving chats by default. With this option, your data is automatically deleted and removed from the system within a certain period — unless OpenAI is legally required to retain it.

ChatGPT has a data retention window period before deleting unsaved chats

Multi-Factor Authentication (MFA)

The MFA feature increases security during the sign-in process, making it more challenging for unauthorized users to access your ChatGPT account. When enabled, MFA means that anyone signing in has to provide multiple forms of identification — typically a password and a confirmation code sent via phone or email.

ChatGPT's security settings are cross-platform and will apply on the website and phone apps

Monitoring and Reporting Features

ChatGPT's usage is actively monitored by OpenAI, with a system in place to respond to any unusual or unauthorized activities. There are also built-in mechanisms to give feedback and report suspicious activities or potential security issues. You can flag inappropriate responses to help reviewers take action or contact support directly and share your safety concerns.

Content Moderation and Filters

The platform has controls to help filter out inappropriate or harmful responses, preventing the dissemination of offensive or biased content. It automatically blocks chats based on prohibited subjects, including illegal activity, hate speech, and violent or adult content.

ChatGPT Safety Risks

Chat Privacy Concerns

The downside to OpenAI monitoring is that ChatGPT's interactions, while encrypted, are not entirely private. Conversations are logged and can be used for model training and improvement. Even if you opt not to save your chat history, it will be stored for a period (as standard practice) before deletion. This raises concerns about the confidentiality of sensitive personal or business information shared during chats.

Data Safety

Because data is logged, there is a risk of your sensitive information getting exposed if data breaches or misuse occur. This could lead to financial, legal, or reputational damage, and criminals may also use the data to launch ransomware attacks, phishing, or identity theft. Because of that, many companies (like Samsung) banned their employees from using ChatGPT.

For example, on March 20, 2023, OpenAI faced an issue with ChatGPT that led to a temporary service outage. During this time, there was a bug that resulted in some users seeing other people's conversation histories. More concerningly, there were fears that payment-related information from ChatGPT-Plus subscribers might have been compromised due to this bug.

Inaccurate or Biased Information

ChatGPT is getting more efficient with time, but the chatbot may inadvertently produce responses containing false or biased information. This phenomenon is being called “hallucination.” If you don't fact-check ChatGPT via independent research, you can end up with incorrect results. This can be dangerous if you’re using the bot to get answers on crucial matters like health or business.

ChatGPT could also be used in media propaganda, manipulating public perception or influencing political decisions by generating content that appears truthful but lacks factual accuracy.

Copyright Infringement

ChatGPT may unintentionally generate content that infringes on existing copyrights, especially in creative fields like writing or coding. It runs on a vast database of past and current information online, so plagiarism may occasionally occur. You could get in trouble or face legal consequences if you unknowingly share or use this content as your own.

Potential for Misuse

Generative AI technology can be exploited for malicious purposes, such as creating phishing scams or malware. Cybercriminals can use the chatbot to impersonate individuals and send convincing fraudulent communications to unsuspecting recipients. This potentially enables them to infiltrate systems and steal information or launch cyber attacks on an entire organization. Students can also use ChatGPT to write essays, which amounts to cheating.

Malicious users still find ways to circumnavigate OpenAI's strict safety policies

Risky Add-Ons and Fake Apps

The popularity of ChatGPT has led to the emergence of many imposter apps (fleeceware) for both phones and computers, as well as risky third-party browser extensions. These counterfeit services may read, write, and access data on your device, posing significant security risks such as data theft and malware distribution. Some people have reportedly paid hefty subscriptions to use these apps, while OpenAI offers Android and iOS apps for free.

Vulnerability to Hacks

OpenAI has implemented robust security protocols to protect ChatGPT users, but as with any online platform, it's not immune to cyber threats. There's a risk of hackers exploiting weaknesses in the system and launching attacks to steal users' data. For example, ChatGPT could be used to construct a convincing email phishing scam. Plus, since even deleted chats are stored for some time, this could put potentially sensitive information at risk.

Prompt injection attacks and training attacks are key concerns for technologies like ChatGPT. Prompt injection attacks occur when specifically crafted inputs trick the AI into unauthorized actions, like getting around content filters or running harmful code. On the other hand, training attacks threaten the model's reliability by sneaking in biased or harmful data during its learning phase — this is often referred to as data poisoning.

Is ChatGPT Safe for Kids?

ChatGPT is designed for users over 13 years old, as mandated by OpenAI's usage policies and privacy laws like the US COPPA. It's advisable for younger children to use ChatGPT under adult supervision to ensure their safety and prevent exposure to inappropriate content. That said, ChatGPT has both educational benefits and risks for children. Some of the ways it can be helpful to kids include:

• Access to education and learning new skills. Under adult supervision, ChatGPT can be a resourceful tool for children's learning, offering answers to various questions and encouraging curiosity. However, the chatbot’s responses should always be taken with caution due to potential inaccuracies or harmful language.
• Improved socialization and self-expression. ChatGPT can serve as a digital companion for socialization, which is especially beneficial for children who might find social interactions challenging, such as those with autism. It can act as a tool to develop language and communication skills.
• Creative and critical thinking development. Interactive storytelling, writing assistance, and exploring new subjects with ChatGPT can stimulate children's creative and critical thinking skills.
• Coding and AI understanding. For children interested in technology, ChatGPT offers an opportunity to understand the basics of coding and AI.

Potential Risks and Concerns for Kids

• Vulnerability to threats online. Children may unintentionally disclose confidential information when chatting to CGPT, for example, where they go to school. If this data ends up in the wrong hands, it could increase the risks of online or offline attacks.
• Exposure to inappropriate content. Despite its built-in safety features, ChatGPT might sometimes generate unsuitable content for children. Therefore, parental supervision is vital to ensure that children's interactions with the AI are safe and appropriate.
• AI plagiarism and cheating on schoolwork. ChatGPT's capability to generate extensive content, including essays and reports, can lead to plagiarism and academic dishonesty. It can be a good idea to chat with children about the values of integrity and original work, and how it helps them to learn.
• Misinformation. Given that ChatGPT's responses are generated from vast but fixed datasets, there's a possibility of encountering outdated, biased, or incorrect information. Kids might not always know that this information needs to be externally confirmed.
• Loss of creativity and self-expression. Over-reliance on AI for creative tasks could hinder a child's natural creative development and problem-solving skills.

How to Use ChatGPT Safely

Alongside the safety features ChatGPT provides, there are some other best practices you can follow to minimize risks when using the platform.

• Only use the official ChatGPT app. Avoid unofficial apps or unauthorized interfaces, as they may not have the same security standards as OpenAI (some can also be intentionally malicious). Only download legitimate apps provided by OpenAI and trustworthy developers, and do thorough research before installing extension add-ons or phone apps from other sources. If you realized you’ve been using a fake app, remove all personal information, uninstall the app, and run an antivirus scan.
• Be cautious with confidential data. Do not share sensitive personal information like passwords, social security numbers, or financial details on ChatGPT. Back-end reviewers can view this data, and you can be a target for cyber attacks if it ends up in the wrong hands. Create a strong password — you can check its strength here.
• Manage your chat history data. You can turn off chat history (a so called “incognito mode”) and delete all stored data for extra privacy. It also means your data won't be used to train the chatbot. To turn on ChatGPT Incognito mode, click the three dots next to your name in the bottom left corner of ChatGPT. Go to Settings, click Data Controls, and toggle Chat history & training off.
  Taking responsibility for your data is the first step to improving your privacy on ChatGPT
• Verify information. Always fact-check content generated by ChatGPT for biases and inaccuracies before using or sharing with others, to avoid causing harm.
• Report vulnerabilities. Flag inappropriate chat responses and report any potential security vulnerabilities that you notice while using ChatGPT. This helps to improve the safety of the chatbot for everyone.

Editor's Note: Transparency is one of our core values at vpnMentor, so you should know we are in the same ownership group as ExpressVPN. However, this does not affect our review process.

FAQs on ChatGPT Safety

Is ChatGPT free?

Yes, ChatGPT is free to use. It also offers a premium version, ChatGPT Plus, which gets you access to Chat GPT-4. Subscribers also get access to additional benefits including early access to new features and faster response times. Both versions of ChatGPT employ the same security measures. However, you should also take individual responsibility to protect your data on the platform.

Does ChatGPT save chats?

ChatGPT does log chat data. Conversations are stored and can be used to train future models or resolve issues/bugs. You can delete or turn off chat history storage, but your data is still stored for a specific duration before it's deleted. This raises privacy concerns when using ChatGPT because OpenAI reviewers could potentially access your chats and use your data.

Is it safe to give my number to ChatGPT?

You no longer need to give your number to register for ChatGPT. That said, you shouldn’t type your number into the chat either. Providing any confidential information on ChatGPT isn’t recommended because OpenAI can share collected data with third parties. Like any other online platform, ChatGPT could also be hacked, exposing your information to cybercriminals.

Is it safe to use the ChatGPT app on an iPhone?

The official ChatGPT app for iOS is generally considered safe to use. There are several safety measures and features to protect users, and OpenAI regularly improves the app's security. However, only download the official ChatGPT app to minimize risks like malware. If you’re not sure whether the app you’re using is official, exercise additional safety measures like using a VPN.

Plus, web browsers and smartphone operating systems like iOS use a security technique known as "sandboxing." It’s a security mechanism used to isolate running programs, making them more secure by restricting their ability to interact with other software or the underlying system. This technique limits the potential damage from a compromised application.

What are the fake ChatGPT apps to avoid?

Any app claiming to be ChatGPT but not published by OpenAI is most likely counterfeit. You'll find many apps with the same or similar names and logos, but these are imposters. If you want to install ChatGPT from your phone's app store, verify the parent company first. Though there are other useful generative AI alternatives to ChatGPT, watch out for apps asking for more data permissions than necessary.

How can I delete my ChatGPT account?

You can submit a request to delete your account through privacy.openai.com or do it yourself. Driven by privacy concerns or safety risks, follow these steps to manually delete your ChatGPT account:

• Sign in to ChatGPT.
• Click on your account icon, located in the bottom left corner of the screen on desktop or in the menu bar on the app.
• Select Settings.
• Navigate to Data controls.
• Click Delete account and confirm your decision.

Winding Up: Is ChatGPT Safe to Use?

ChatGPT is a transformative AI tool we can safely leverage for most aspects of our daily lives, whether it’s for learning, creativity, or professional use. As with any new technology though, it also has its fair share of security challenges that must be mitigated. To avoid the worst risks, make sure you’re using the platform’s available security measures and always take extra steps to protect yourself online.