How to prevent original ideas from leaking out in automatic program generation functions

Abstract

As generative AI and automated programming that does not require human coding become mainstream, code generated by automated program generation functions using similar natural language input will become equivalent code and will be more likely to be imitated by others. The system records the sentences or phrases entered into the automatic program generation function, and when an equivalent sentence or phrase is subsequently entered, it disables program execution for the latter input, thereby preventing the generation of programs generated by the input of the equivalent sentence or phrase.

Description

[Technical Field]

[0001] The present invention relates to an automatic program generation function that automatically creates software, so-called automatic programming, and to an environment that executes multiple automatic programs, i.e., a technology that is effective when applied to the automatic program generation function. [Background technology]

[0002] As generative AI becomes more widely used and automatic program generation functions that do not require human programming—known as automatic programming—become mainstream, it is believed that programs will be generated by artificial intelligence using natural language processing, reducing the need for human programming. Currently, automatic programming generates programs using associations between natural language and code created by humans. The programming code for natural language is associated with a database associated with the automatic program generation function or with a program that has the automatic program generation function. Corresponding code is added for new words and sentences. This code is inserted into the program with the automatic program generation function or into a shared database. This results in common learning and growth, regardless of the natural language input by each user, and all users can utilize the learned code and natural language association. This structure, which has natural language-compatible code within a database, allows multiple databases to be sequentially synchronized and use the same database. Furthermore, a structure with natural language-compatible code within an automatic program generation function allows the same automatic program generator to be maintained by adding the learning results to all automatic program generation functions, recompiling them, and sequentially synchronizing them. This allows all users to enjoy the same learning results.

[0003] On the other hand, a disadvantage of this uniform environment for automatic program generation is that using the same natural language scenario can result in the generation of identical programs. Even if a unique idea is realized and an application software with added value is developed, similar applications can be developed by providing similar natural language sentences, resulting in a loss of competitiveness. Furthermore, when automatic program generation functions have learning capabilities like generative AI, a program created from a natural language sentence provided to one automatic program generation function can be shared by all automatic program generation functions, potentially resulting in the loss or leakage of the originality of the idea. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Patent Publication No. 09-073306: A method for inputting data into a database that is closely related to program input, by making the input method for creating a program closer to natural language (several types are proposed) [Patent Document 2] Patent Publication No. 06-214776: Automatically generates target program source code by inputting software specifications written in natural language [Patent Document 3] Patent 7064680: A natural language processing method has been added to JP Patent Publication No. 06-214776, and the corresponding program code has been stored in a database. [Patent Document 4] Patent 6958884: The content is the same as JP 06-21477, but with the addition of a conversation reception function. [Patent Document 5] Patent Application No. 2018-212991: Method for generating feedback information associated with input content and providing it to a user terminal [Patent Document 6] Patent application 2012-88951: Method for reflecting modifications made to a newly automatically generated program when modifications are made to a program generated by automatic programming [Patent Document 7] Patent application No. 62-39145: Artificial intelligence that automatically creates programs by providing natural language to the artificial intelligence and adding minimal human support [Patent Document 8] Patent 7573141 Preventing idea leakage during automatic programming [Patent Document 9] Patent 7575736 Preventing idea leakage during automatic programming Summary of the Invention [Problem to be solved by the invention]

[0005] In systems or programs with automatic program generation functions, including generative AI, when the same automatic program generation function is used in different environments, inputting any similar natural language sentence will result in the generation of similar program code. This means that the original ideas in the input natural language sentences can leak out through the generated similar program code. For example, this means that programs with similar functions can be easily generated by writing functions from existing application functions, and there is a possibility that programs such as cryptographic technologies can be created in the same way and then decrypted. [Means for solving the problem]

[0006] A system with an automatic program generation function or the program generation function itself memorizes the input sentence or phrase, and when a similar sentence or phrase is input from the second time onwards, it generates an inexecutable program, making it impossible to run the program after it has been generated for the input of a similar sentence or phrase, thereby preventing the leakage of ideas. [Brief explanation of the drawings]

[0007] [Figure 1]Examples of general AI flow, including generative AI, automatic program generation functions, and so-called automatic programming natural language processing, which divides the given content and extracts code from a database according to the content, and examples of code extraction when using generative AI. [Figure 2] An example of a program generation function that has the content of an input sentence or phrase and can optionally set whether to allow the program to be executed if the content is equivalent. [Figure 3] Example of a flow to allow / disallow execution within a program when someone else enters the same information after you. [Figure 4] Example of execution permission flow for equivalent programs that set whether they can be executed when generating a program from text input DETAILED DESCRIPTION OF THE INVENTION

[0008] Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. However, the present invention can be implemented in many different forms and should not be interpreted as being limited to the description of the present embodiment. Note that the same elements will be assigned the same numbers throughout the embodiments.

[0009] In the following embodiments, a method or a system will be mainly described. However, as will be apparent to those skilled in the art, the present invention can also be implemented as a program usable by a computer. Therefore, the present invention can be embodied as hardware, software, or a combination of software and hardware. The program can be recorded on any computer-readable medium, such as a hard disk, a DVD-ROM, an optical storage device, a magnetic storage device, or a semiconductor storage device.

[0010] In addition, the following embodiments may use general computer systems, etc. Computer systems, network systems, and systems for artificial intelligence that can be used in the embodiments include hardware resources typically found in computer systems, network systems, and systems for artificial intelligence, such as a central processing unit (CPU), a main memory (RAM), a non-volatile memory (ROM), a coprocessor, a graphics accelerator (GPU), semiconductors for AI, cache memory, an input / output control device (I / O), semiconductors for artificial intelligence functions, and flash memory. They may also include external storage devices such as hard disk drives and flash memory, and communication means connectable to networks such as the Internet. Computer systems include various types of computers, such as personal computers, workstations, mainframe computers, and servers, as well as cloud-based network-connected computers, tablet computers, various portable information terminals such as smartphones, and systems that implement artificial intelligence.

[0011] The present invention aims to prevent the output of similar program code even when highly unique input content is used, by adding program code generated by an automatic program generation function to the automatic program generation function, or so-called automatic programming, so that the learning and growing generation AI and automatic program generation function can output similar programming code when similar natural language sentences are input as a result of the growth of the automatic program generation function. In other words, the purpose is to prevent the violation of confidentiality and the leakage of uniqueness.

[0012] Generative AI and other systems with automatic programming capabilities generally divide the input text into phrases, as shown in Figure 1, extract program elements corresponding to those phrases from a database, etc., and combine them arbitrarily or according to a set of rules established by the system to generate a program. The results are then further combined to generate a program that meets the purpose.

[0013] As shown in Figure 2, the present invention comprises an automatic program generation function that stores input sentences or phrases separated from the input sentences, and a function that, if an input equivalent to these input sentences or phrases is received after storage, sets a permission flag associated with a user ID to allow or disallow program execution within the program when generating a program corresponding to the input sentence or phrase. Also, by optionally setting a permission period for the user ID, it becomes possible to set a period during which the program can or cannot be executed.

[0014] Figure 3 shows an example of a flow in which a user can set whether to allow or disallow execution of a program in a program when it is generated if the user inputs their own sentence, or a segmented sentence that exists independently as content, or a combination of such sentences after the user's own sentence. If the user who first inputs the sentence sets it as "allowed," then even if another person inputs the same content, that other person can set the program as executable or inexecutable when automatically generating the program. For example, if the setting is set to "disallowed," and similar content has already been saved, an inexecutable program will be generated when another person automatically generates the program, even if they input the same content.

[0015] Figure 4 shows an example of a simplified flow for the program generation prevention function. To automatically generate a program, a sentence describing the desired function is input. In generative AI and automatic programming, the input sentence is typically processed using natural language processing, for example, by dividing it into sentences or phrases that make up unit functions, and then by content. The input sentence content or the divided phrase content is checked to see if a sentence or divided phrase with equivalent content exists in a data storage device, such as a database, where the input was saved before. In other words, a comparison is made with the saved content. If a sentence or divided phrase with equivalent content does not exist in the saved content, the content is saved as new content along with the user ID of the person who entered the content. The automatic program generation function then generates a program based on the input content. The resulting program is generated. If the saved data already contains equivalent content, the program execution permission flag associated with the user ID who first entered the saved data is checked. If the program execution permission flag is set to "allow," the automatic program generation function sets program execution permission within the program based on the input content and then automatically generates the program. If the program execution permission flag is not permitted, the program automatic generation function will set program execution not permitted within the program based on the input content and then automatically generate the program. As a result, the program will be generated by the program automatic generation function. If the permission flag is not permitted, the program will be not permitted and the user will not be able to execute it. As a result, it is possible to prevent the generation of an equivalent program even if there is equivalent input. Furthermore, if the user who entered the first sentence gives permission to the user ID that enters the second and subsequent equivalent sentences, it will be possible to generate an executable program. Furthermore, if the user who entered the first sentence gives permission to execute all subsequent inputs, a normal executable program will be generated. [Industrial Applicability]

[0016] It can be applied to generative AI, which builds unique application software for each individual or company, preventing it from being leaked to others.

[0017] In addition, the uniquely generated program codes can be used on sites that sell them for a fee, and by applying this to a generation AI that provides codes for pre-created functions only to registered users, it becomes possible to give permission for unique know-how to be used only by specific users.

[0018] Furthermore, by adding settings such as a period during which the program can be executed, it is possible to generate programs that can be provided as programs that can be executed only during the licensed period or programs that cannot be executed until a certain period has elapsed.

Claims

1. A system equipped with a generation AI or automatic program generation function that generates a program containing an error when a sentence or phrase combination with the same content as the first sentence or phrase combination is inputted second or later.

2. The system according to claim 1, comprising a generation AI or automatic program generation function that generates an error-free program when the first inputter gives permission to other people who input data after that.

Images