With the advent of services based on modern generative neural networks, which have proven to be extremely effective in various fields and very useful for people, new security risks have emerged. One of them is an advanced approach to password disclosure called PassGAN. Using the latest AI model, 51% of common passwords can be successfully compromised in less than one minute, with 71% of the passwords easily cracked within a day. A terrifyingly serious problem for society in the malicious use of AI was demonstrated by Home Security Heroes as part of a large-scale special study.
The table shows the number of characters and the speed of hacking by artificial intelligence – only unique complex passwords count for decades
PassGAN was used on a large database of over 15,600,000 different passwords. The results are not just stunning – they literally make you panic. In less than a month, the presented AI will take 81% of ordinary passwords to crack. The speed depends on the volume of characters and the complexity of their combination, but the main concern is not just performing traditional brute force by generating samples, but the ability of GAN projects to autonomously learn from Internet database leaks, analyze real character sets, and generally approach the matter on a larger scale. . For example, it takes only 7 hours to crack an atypical 8-element medium-level password with a common combination of lowercase and uppercase letters, numbers, and PassGAN symbols.
The table shows the number of characters and the speed of hacking by artificial intelligence – only unique complex passwords count for decades
The Home Security Heroes official website offers a tool that allows you to check your passwords for resistance against powerful artificial intelligence algorithms. Experts assure that no transmitted information is stored on the company’s servers and is not sent to third parties. However, due to excessive caution, it is safer not to enter any of the actual passwords. If you want to try an AI password analyzer, it’s better to introduce a theoretical random selection of characters.
If you are interested in learning more about the processes of how the PassGAN neural network works from the inside, Home Security Heroes has prepared a detailed report .
Tips on how to protect yourself from AI password cracking :
- be sure to stick to passwords of at least 12 characters with lowercase and uppercase letters and with the addition of numbers (or individual characters);
- activate two-factor authentication technology (not via SMS code, if possible);
- use automatically generated passwords (they are always complex and least susceptible to quick hacking);
- regularly monitor passwords in case of leakage (Apple and Google have such systems) and update them, especially for confidential accounts;
- refrain from connecting to public Wi-Fi networks.
