Why you shouldn't create your passwords with ChatGPT and other AIs?

In recent years, the artificial intelligence applied to the domestic and professional spheres It has become an everyday tool. Many people already use assistants like ChatGPT to compose texts, analyze information, generate ideas, and even for digital security-related tasks, such as generating passwords. However, Is it really safe to let an AI protect your personal or work accounts?

Although it may seem convenient and even futuristic to ask ChatGPT or similar models to invent a “secure” password for you, the reality is quite different: There are hidden risks and technical limitations that make this practice not only inadvisablebut can leave you exposed to cyberattacks or unexpected leaks.

Below, we break down in great detail the Reasons why you shouldn't use ChatGPT (or other similar AI) to create your passwords and what alternatives you have at your disposal to keep your accounts safe.

Top reasons to avoid using ChatGPT for password creation

The popularization of ChatGPT and other generative AIs has led to more and more users turning to them for tasks as varied as text generation and problem-solving. However, when it comes to creating passwords, the consensus among cybersecurity experts is clear: it's not a reliable option.

• Pattern predictionAI systems, no matter how sophisticated, don't generate truly random combinations. They tend to repeat certain patterns or sequences that cybercriminals can exploit.
• Non-unique passwordsChatGPT may suggest the same password to different users when given similar instructions, increasing the risk of multiple accounts sharing the same vulnerable password.
• Cloud data storage: Everything you enter into ChatGPT is logged on the company's servers, making those passwords an attractive target for potential security breaches or leaks. Learn how to manage your passwords on Google to avoid similar risks.
• Risk of unsafe suggestionsIn numerous tests, ChatGPT has generated passwords with common words disguised as symbols, predictable phrases, or easy-to-guess number patterns.

These reasons aren't just theoretical. A recent report by Kaspersky and analyses by various experts have tested the robustness of thousands of AI-generated passwords, including ChatGPT, Meta's Llama, and DeepSeek, yielding alarming results in both password predictability and resistance to cyberattacks.

Why don't AI models generate truly strong passwords?

The essential difference between a professional password generator and a conversational AI lies in the concept of randomness and uniqueness. Specialized generators They use cryptographically secure algorithmswhile language models like ChatGPT They operate on statistical patterns learned from large amounts of text. This causes:

• Pattern imitation: AIs tend to create keys that, although they appear to be random, follow regularities and structures recognizable to someone who knows how these models work.
• Inability to generate complete randomness: LLMs (Large Language Models) like ChatGPT do not use randomization functions typical of cryptography, but rather predict the next letter or character based on what is most likely based on their learning.
• Use of common words and sequences: It is common for some generated passwords to include dictionary words camouflaged with numbers or symbols (e.g., “B@n@n@7,” “S1mP1eL1on”), or combinations such as “123” or the word “password,” which is highly discouraged.

This means that, in the event of a massive brute-force attack by hackers, attempts can focus on the most common patterns instead of being completely random, greatly speeding up the possibility of cracking your credentials.

Recent study results: How vulnerable are these AI-generated keys?

Concerns about the insecurity of passwords created by artificial intelligence are not just a theoretical assessment. Kaspersky's research team conducted a study with thousands of keys generated by ChatGPT, Llama and DeepSeek, checking its robustness and ability to survive automatic attacks.

• In the case of DeepSeek, a 88% of passwords generated were deciphered in brute force tests.
• Calls was close, with a 87% of keys dropped in less than an hour.
• ChatGPT offered better resultsBut still, 33% of the passwords created were considered insecure by professional standards.

Beyond the success rate, the researchers found that Many of the characters, numbers and symbols were frequently repeated in the passwords proposed by AI, which reduces their effectiveness against automated attacks. For example, certain symbols and letters appeared systematically, and sometimes the templates did not include the recommended minimum length of 12 characters or all the required character types.

The danger of entrusting sensitive information to ChatGPT

Everything you enter into ChatGPT is stored on the company's servers, which can compromise the confidentiality of your data. This storage could be vulnerable to security breaches. or misuse of information by OpenAI.

• Passwords can be exposed in a data breach.
• OpenAI may analyze, share, or transfer your data to third parties. in certain circumstances, increasing potential risks.
• No confidentiality is guaranteed: Unless you configure all privacy options and disable chat history, your data may be stored and used.

Therefore, experts recommend never enter critical data on AI platforms, especially passwords and other sensitive data, to avoid potential vulnerabilities.

Why might passwords be duplicated between multiple users?

Asking ChatGPT for passwords with generic commands can result in other users receiving the exact same suggestions. This happens because:

• Similar prompts produce equal responses: If many people ask “give me 10 strong passwords,” the answers will be similar or the same.
• Limitations in randomness: There is no guarantee of exclusivity, no matter how complex the key may seem.

The consequence is that multiple people could use the same password, making it easier for a cybercriminal to try that combination on different services and successfully access multiple accounts. Passwords must be unique and non-transferable., something that current AI cannot guarantee.

Real-life examples of insecure passwords created by AI

Studies show Examples of passwords generated by ChatGPT, Llama, and DeepSeek, with patterns like:

• B@n@n@7
• S1mP1eL1on
• Password
• P@ssw0rd!23
• qLUx@^9Wp#YZ
• LU#@^9WpYqxZ
• YLp^9W#qX@zv

Many include dictionary words camouflaged with symbols, making them vulnerable to dictionary attacksOthers, although seemingly random, repeat symbols and patterns, making them easier to decipher by automation and reducing their security.

Differences between a password generator and ChatGPT

Compared to general AI models, specific solutions for creating strong passwords are:

• Password generators: They use cryptographically secure algorithms that produce unpredictable strings of configurable length and with different types of characters, eliminating repetitive patterns.
• Password managersIn addition to storing credentials in encrypted boxes, many include generators that produce very strong, random keys.

It is always recommended to use these solutions because:

• They allow you to create long, unique, and random passwords without relying on memory or insecure combinations.
• They facilitate auto-completion, synchronization, and leak alerts.
• End-to-end encryption prevents third parties from accessing your keys.

What privacy risks does using ChatGPT to create passwords pose?

One important aspect to consider is that Everything you enter into ChatGPT is stored on their serversThis can pose a risk if your data, including passwords, is exposed in breaches or used without your consent.

• Risk of attacks on servers: A breach can leak your passwords and sensitive data.
• Possibility of sharing data with third parties: OpenAI may share information in some cases, increasing exposure.
• Lack of guaranteed confidentiality: It is not recommended to enter sensitive data on the platform.

For this reason, all experts advise Avoid entering important or confidential data into ChatGPT or similar platforms.

Practical tips and secure alternatives to protect your accounts

To create and maintain strong and unique passwords, follow these recommendations:

• Use professional password generators, integrated into managers such as Keeper, Dashlane, 1Password, Bitwarden or Kaspersky Password Manager.
• Choose long passwords (at least 16 characters), combining uppercase, lowercase, numbers and symbols, without dictionary words.
• Do not reuse passwords in different services.
• Store your passwords in trusted managers and avoid unencrypted notes, emails, or chats.
• Activate two-step verification whenever you can.
• Do not share or enter your passwords on AI platforms or unknown sites..

If you use passphrases, combine several random words and different characters to increase their security and difficulty of cracking.

The convenience of asking ChatGPT or another AI to generate a password does not compensate for the risks associated with predictability, repetition, and storage on external servers. Nowadays, there are reliable, easy-to-use, encrypted alternatives that offer real protection for your digital accounts.