Brute-force resistant encryption method

Abstract

The present invention relates to a novel method for protecting encrypted information against decryption attempts by quantum artificial intelligence. The invention is based on the prior modification of the encrypted information so that a decryption attempt results in an incorrect solution or that several decryption attempts result in solutions plausible to quantum artificial intelligence. Thus, even the most powerful artificial intelligence, even one powered by quantum computing, can generate either an incorrect solution or a large number of plausible solutions without being able to determine which one is correct. The method described in the present invention represents a significant advance in the field of encryption, rendering data impenetrable to decryption attempts by the most sophisticated quantum artificial intelligence systems. Abstract figure: Fig. 1

Description

Title of the invention: Brute-force resistant encryption method Technical field

[0001] The invention is a means of securing a set of data that resists attempts at decryption by brute force technique. Prior art

[0002] We know of widely used encryption algorithms such as RSA and elliptic curve cryptography (ECC), which quantum computers could potentially break, making digital identities vulnerable to theft, fraud and exploitation, and more generally putting almost all Internet security at risk. Technical problem

[0003] The importance of identification will only increase with the advent of quantum computing and advances in artificial intelligence, which will make it possible to find all encryption keys in a very short time, through brute-force attacks.

[0004] Various solutions are being studied such as replacing cryptographic algorithms with algorithms resistant to quantum computing, improving key lengths or using post-quantum cryptography, but the transition to these methods may take time.

[0005] The only known and practiced method today against brute force consists of limiting the number of decryption attempts for documents encrypted by a person, which reinforces the interest in strong identification methods for people.

[0006] The objective of the present invention is to resist attempts to decrypt files by brute force. Brief description of the drawings

[0007] The invention will be well understood, and other objects, advantages and features thereof will become more apparent from the reading of the description which follows, which is illustrated by [Fig.1].

[0008] [Fig. 1] is a diagram showing a procedure for reading and verifying a protected digital file using a method according to the invention. The receiver reads the message, called the "mes sage-signifiant," which is a signed barcode containing the address of a digital file, called the "fichier-unintelligible," on a data server and an integrity control element for this "fichier-unintelligible." It verifies this barcode by its electronic signature, contacting a signature server if necessary, and then transmits it. the address of the unintelligible file to the data server containing it. It receives in return the decrypted digital file that the data server obtained by combining the unintelligible file and the meaningful message, and verifies it with an integrity check element. Statement of the invention

[0009] The invention is a method for encrypting a set of data comprising the following steps: 1. Modify, using an algorithm called a modification algorithm, all or part of said dataset to obtain a file called an unintelligible file; such modifications may involve data deletion and, optionally, encryption using a conventional method. 2. Calculate a message called a meaningful message containing an indication of a means to decrypt said unintelligible file, this means comprising an algorithm called a decryption algorithm, characterized in that: 1. The modification performed by said modification algorithm is such that a plurality of decryption attempts carried out without said decryption algorithm result in an incorrect solution or different plausible solutions, 2. and that the said wise significator allows the validity or invalidity of a solution obtained by an attempt at deciphering to be determined. Detailed description of the invention

[0010] Decrypting a file encrypted by the brute force method consists of imagining a very large number of decryption keys and trying to decrypt the file in question by many successive attempts made with each of these imagined decryption keys.

[0011] Decryption is considered a success when a plausible solution has been obtained, for example an identity photograph if it is such a photo encrypted, or the text of a contract if it is a contract.

[0012] If the decryption results in an incorrect solution, it is a failure.

[0013] If it results in a plurality of plausible solutions, the best quantum artificial intelligence cannot determine which is the correct solution.

[0014] The present invention is based on the principle of making a prior modification of the encrypted information such that an attempt at decryption leads to an incorrect solution, or that several attempts at decryption lead to plausible solutions between which the decryption method cannot choose which is correct.

[0015] Advantageously, more than a hundred attempts at decryption should lead to different plausible solutions, but an improved version of the present method can also be such that several thousand attempts at decryption lead to different plausible solutions.

[0016] For maximum efficiency, the modification of the dataset performed by said modification algorithm focuses on data essential to the use of the dataset in question, that is, on data without which the dataset is no longer useful for the intended purpose. It is advantageous to use artificial intelligence to determine what type of dataset it is and, consequently, to determine what modifications need to be made to it.

[0017] In a preferred version, the modification algorithm involves, in a first step, analyzing all or part of the dataset and, in a second step, defining the modifications to be made to render it unintelligible and to ensure that several decryption attempts result either in an incorrect solution or in a plurality of plausible solutions. Advantageously, this operation is performed by an artificial intelligence specifically trained for this purpose.

[0018] Here are some examples: 1- In the case where the data set is the description of a security marking characterized by a series of characteristic pixels of color, and / or brightness, and / or light reflection varying according to the point of view, the modification algorithm can be very simple since any modification or deletion of data is essential, and any solution can be plausible; 2- in the case where the dataset is a descriptive file of an object or a person, the modification algorithm advantageously includes the modification or deletion of dimensional dimensions of the object or person considered; 3- in the case where the dataset includes an illustration such as a drawing, a photograph or a video of an object or a person, the modification algorithm advantageously includes distortions or deletions of all or parts of the illustration in question; 4- In the case where the dataset includes text such as the description of a method or device, a contract, an order, a report, a press article, the modification algorithm advantageously includes modifications or deletions of mentions essential to the interpretation or understanding of the text in question.

[0019] In the case where the dataset is a key controlling access to a physical or digital domain, a solution is advantageously considered plausible as soon as it opens access, even if the domain is not the right one.

[0020] It is advantageous that the physical or digital domains to which access can be given by brute force resembles that which the original dataset provides access to so that the corresponding solutions are considered plausible.

[0021] The same unintelligible file can be combined with several different meaningful messages. In combination with the same unintelligible file, different so-called secondary meaningful messages can, for example, produce an identity card or passport, a visa, a driver's license, a social security card, a disability certificate, a sick leave certificate, an insurance certificate, a power of attorney, a membership card, an invitation, a train ticket, a payment card, etc.

[0022] It is advantageous that a primary meaningful message, for example issued by a public authority or a trusted third party, can be used by third parties such as banks, transport companies, event organizers, administrations and private companies and even individuals to generate such secondary meaningful messages.

[0023] To enable the emission of secondary meaningful messages, it is advantageous that the so-called primary meaningful message that enabled their generation includes a part easily identified as having to be reproduced in a secondary meaningful message so that the latter can be combined with the unintelligible file associated with the primary meaningful message, or that it includes a reference to which the secondary meaningful message can be associated.

[0024] The decryption algorithm is advantageously provided with an algorithm reference and the message-signifier includes a means of obtaining it, for example its reference in a database.

[0025] It is desirable that the means of obtaining the decryption algorithm should only be provided to a person or computer responsible for decryption after verification that the person or computer is indeed authorized to decrypt the data set.

[0026] Strong identification is therefore desirable. It makes it possible on the one hand to determine which part of the population or computers can decipher the unintelligible message, but also to limit the number of successive attempts to decipher the unintelligible message, which would not be possible without strong identification, since a fraudster could use several different identities, for example several IP addresses, to make a large number of successive attempts in a limited time.

[0027] The meaningful message and the unintelligible file are advantageously one and / or the other provided with a means of controlling their own integrity, such as a hash, a checksum for example, and / or a digital signature allowing their author to be authenticated and their non-repudiation to be guaranteed.

[0028] The meaningful message and the unintelligible file can each be equipped with such means of control concerning respectively the unintelligible file and the meaningful message.

[0029] The message-signifier may take the form of a string of characters, a barcode, or an internet link. It may include references or links to the unintelligible file, including multiple links to distinct parts of that file that may be stored at different addresses.

[0030] The unintelligible file must not contain any information that would allow it to be linked to a meaningful message.

[0031] An additional means of security may be the opposite of not including any reference or link allowing the unintelligible file to be found.

[0032] A secure implementation can therefore be to send to the person authorized to decrypt the unintelligible file these three elements, namely the unintelligible file, the meaningful message, and the means of accessing the unintelligible file separately, by different means of communication, and without any of these three elements allowing it to be brought closer to one or the other of the other two.

[0033] In the application to the identification of persons, it is recommended for example to send to the holder of a digital identity the unintelligible file containing unintelligible data from his biometric characteristics either by an Internet link allowing it to be downloaded, or by an email, and to send the meaningful message by text message or by postal mail or by hand delivery.

[0034] The signifying message may also contain non-confidential information in plain text, advantageously signed, such as the title of the protected information or certain of its characteristics. This allows identities or certificates to be signed and communicated to third parties, who may, where appropriate, obtain additional information, such as an identity photograph or a certified copy of a contract, by submitting the signifying message to the appropriate server.

[0035] In a more secure version, the file can only be decrypted by a secure computer. The decryption algorithms can then remain secret.

[0036] In this more secure embodiment, the meaningful message or the unintelligible file is advantageously kept confidential by a secure computer, which uses it to combine it respectively with the unintelligible file or the meaningful message as appropriate.

[0037] In an even more secure version, the result of combining the unintelligible file and the meaningful message is also secretly stored by the secure computer. This prevents the dissemination of protected data sets according to the invention. The user only needs to send what they believe to be the original dataset, or what that dataset represents, so that the secure computer can respond with yes or no as to whether what it sent is authentic or not.

[0038] Except when it is only a matter of issuing encrypted data sets according to the present invention to reconstruct a not-so-secret file such as an identity photo which one seeks to prevent from being used to create databases enabling unauthorized persons to perform facial recognition, this more secure implementation is especially recommended for the protection of other biometric characteristics such as fingerprints or DNA signatures, or for highly confidential documents.

[0039] The present invention is also a set of data protected by an encryption method according to one of the methods described above.

[0040] It is also a method of decryption described above of such a data set Applications

[0041] The methods described above apply to all types of files requiring confidentiality. They can be applied to all forms of files representing images, music, videos, software, graphics, engineering plans, texts such as medical prescriptions, medical examination reports, expert opinions, contracts, payment orders, etc.

[0042] The main applications are - digital identity, the holder of which may only keep a signed meaningful message, in printed or electronic form, to identify himself, - medical prescriptions, certificates and attestations of all kinds, contracts, purchase orders, invoices, etc. which can be easily authenticated by their recipients, - music or video on demand services, newspapers and magazines, video games, means of controlling connected objects, - and generally the securing of any type of large or sensitive data.

Claims

Claims

1. Method for encrypting a data set comprising the following steps:

1. modifying by an algorithm called a modification algorithm all or part of said data set to obtain a file called an unintelligible-file, such modifications possibly including deletions of data, and optionally encryption by a conventional method, 2. calculating a message called a signifying-message containing the indication of a means of decrypting said unintelligible-file, this means comprising an algorithm called a decryption algorithm, characterized in that:

1. the modification carried out by said modification algorithm is such that a plurality of decryption attempts carried out without having said decryption algorithm result in an incorrect solution or in different plausible solutions, 2.and that said message-signifier makes it possible to determine the validity or invalidity of a solution obtained by an attempt at decryption.

2. Encryption method according to claim 1 characterized in that more than a hundred decryption attempts result in different plausible solutions.

3. Encryption method according to claim 1 or claim 2 characterized in that said modification algorithm relates to data essential to the use of said data set.

4. Encryption method according to any one of the preceding claims, characterized in that said modification algorithm comprises in a first step the analysis of all or part of said data set to define in a second step what said modifications are to be made to it to make it unintelligible.

5. Encryption method according to claim 3 or claim 4 characterized in that: - in the case where said data set is a descriptive file of an object or a person, said modification algorithm comprises the modification or deletion of dimensional values ​​of said object or said person, - in the case where said data set comprises an illustration such as a drawing, a photograph or a video of an object or a person, said modification algorithm includes distortions or deletions of all or parts of said illustration, - and in the case where said data set includes a text such as the description of a method or device, a contract, an order, a report, a press article, said modification algorithm includes modifications or deletions of information essential to the interpretation or understanding of said text.

Images