Sharing of artificial intelligence models in a clean room with controlled data access

Abstract

A secure clean room environment for sharing artificial intelligence (Al) models between multiple parties uses access control mechanisms and role-based access control for governing access to the Al models within the clean room environment. Isolated sandbox environments within the clean room environment allow authorized parties to test and use the Al models, monitor and log all activities performed within the clean room environment, and enforce usage controls on the Al models. The clean room leverages a retrieval-augmented generation (RAG) framework, multi-agent architecture, and confidential computing to enable safe, purpose-specific machine learning (ML) model sharing. A centralized knowledge graph and confidential virtual machines (VMs) create isolated environments where partners can securely access and query ML models.

Description

Attorney Docket No. RAMP-00314-WOSHARING OF ARTIFICIAL INTELLIGENCE MODELS IN A CLEAN ROOM WITH CONTROLLED DATA ACCESS CROSS-REFERENCE TO RELATED APPLICATIONS

[0001] This application claims the benefit of US Provisional Patent Application No. 63 / 720,421, filed November 14, 2024, and US Provisional Patent Application No. 63 / 760,992, filed February 20, 2025. Such applications are incorporated by reference herein.BACKGROUND OF THE INVENTION

[0002] With the increasing collaboration in artificial intelligence (Al) development, the need to share Al models securely among partners has become critical. Traditional methods of sharing models pose risks such as unauthorized usage, data breaches, and infringement of intellectual property rights. Raw, ungoverned information cannot be left unprotected. Although vector databases enable structured Al consumption, they lack a control framework to define what data Al models can retrieve based on policies, permissions, and regulatory constraints; how responses are filtered before returning to end users; and what data-level restrictions should be enforced to prevent unauthorized or unintended information leakage.

[0003] A data clean room is a secure, privacy-preserving environment within a cloud computing network that allows multiple organizations to combine and analyze their respective data sets without directly sharing raw data with each other. Such environments use advanced encryption and anonymization techniques to ensure that participating parties can only access aggregated insights and analysis results, rather than individual-level data. This technology enables companies to collaborate on data projects, match customer databases, measure advertising effectiveness, or conduct joint research while maintaining strict compliance with data privacy regulations and protecting sensitive business information. The clean room acts as a neutral intermediary, applying predefined rules and controls to prevent the extraction or reverse engineering of original data sets while still allowing valuable insights to be derived from the combined data.

[0004] The inventors hereof have recognized that it would be desirable to develop systems and methods for sharing Al models by leveraging secure clean room environments.Attorney Docket No. RAMP-00314-WOSUMMARY OF THE INVENTION

[0005] The invention is directed to a system and method for providing a secure clean room environment for sharing Al models between multiple parties, using access control mechanisms and role-based access control for governing access to the Al models within the clean room environment. Isolated sandbox environments within the clean room environment allow authorized parties to test and use the Al models, monitor and log all activities performed within the clean room environment, and enforce usage controls on the Al models.

[0006] In certain embodiments, the secure clean room implementation leverages a Retrieval- Augmented Generation (RAG) framework, multi-agent architecture, and confidential computing to enable safe, purpose-specific machine learning (ML) model sharing. A centralized knowledge graph and confidential virtual machines (VMs) create isolated environments where partners can securely access and query ML models.

[0007] In certain embodiments, multi-agent RAG components manage layered security. A retrieval module ensures only authorized, role-specific data access, while a generation module, operating within the VMs, filters outputs based on compliance rules and user purpose. Within these secure VMs, confidential computing technology encrypts data and computation, preventing unauthorized access throughout processing, even from cloud administrators.Automated compliance agents and real-time policy monitoring enforce regulatory adherence, logging all interactions for audit trails. Only purpose-filtered, encrypted outputs are allowed to exit the clean room, ensuring data privacy and control across all stages.

[0008] In addition, in certain embodiments the invention comprises a control layer that provides "guardrails" to implement policy-driven Al access control to regulate retrieval and response outputs.

[0009] This integrated clean room architecture establishes a highly secure, compliant environment, fostering trusted Al model collaboration. The architecture thereby ensures that Al models are used only for their intended purposes under strict security protocols. The result is an Al model sharing environment that exhibits data privacy, security, and usage compliance. The invention provides a controlled environment where Al models can be accessed and utilized.

[0010] The invention provides a secure clean room environment that allows for controlled sharing and usage of Al models. In various embodiments, there are a number of features of the specialized Al model-sharing clean room.Attorney Docket No. RAMP-00314-WO

[0011] One feature of the secure Al model-sharing environment is access control, which provides strict authentication and authorization mechanisms to ensure that only approved partners can access the clean room.

[0012] Another feature is continuous monitoring and logging of all activities within the clean room to ensure compliance with usage policies.

[0013] Another feature is end-to-end encryption of Al models and data to prevent unauthorized access.

[0014] Another feature is sandboxing, that is, isolated environments where partners can test and use Al models without risking the integrity of the original models.

[0015] Another feature is compliance management, which includes tools to enforce and verify compliance with data privacy regulations and contractual agreements.

[0016] Another feature is usage control, which enable usage of an Al model during a specified time period, at a certain frequency limit (per unit time or cumulative), and in a granular specification of allowed or disallowed use-cases.

[0017] As can be seen from these features, data clean rooms adapted specifically for Al model sharing provide a secure and controlled environment that addresses several key challenges in Al model sharing. In their traditional form, data clean rooms allow multiple parties to analyze combined datasets without exposing raw data. When adapted for Al model sharing according to embodiments of the invention as disclosed herein, these environments create a secure intermediary space where model owners can make their Al models available to partners while maintaining strict control over their intellectual property and usage.

[0018] In a non-limiting example, a company that has developed a sophisticated fraud detection Al model might want to license it to financial institutions without risking the model being copied or reverse engineered. Using a data clean room environment, the model owner can make the model accessible through an API interface within the clean room, where partner banks can send their transaction data for analysis and receive fraud detection results. The clean room enforces strict controls, such that the partner banks can't extract or copy the model itself, but can only use it for approved purposes within defined limits. The environment might allow the banks to test the model's performance on sample datasets, integrate it with their systems through secured APIs, and receive aggregated performance metrics, all while preventing access to the model's underlying architecture and weights.Attorney Docket No. RAMP-00314-WO

[0019] Another practical application for certain embodiments of the invention described herein is in collaborative Al development across organizations. Consider pharmaceutical companies working together on drug discovery Al models. Through a clean room environment, each company can contribute their proprietary models for specific aspects of drug discovery (such as protein folding or binding affinity prediction) without exposing their underlying intellectual property. The clean room can enforce usage quotas, track how models are being used, ensure data privacy compliance, and even manage billing based on usage. Partners can use these models through controlled interfaces that limit what information they can extract, while still allowing them to build valuable applications on top of the shared models.

[0020] A key innovation in each of these examples is that data clean rooms are being adapted from their original purpose of secure data sharing to handle the unique requirements of Al model sharing, which may include, for example, model versioning, usage monitoring, performance tracking, and the prevention of model extraction attacks. This enables a new paradigm of collaborative Al development where organizations can monetize their Al models while maintaining control over their intellectual property to prevent infringement or other unauthorized use.

[0021] In one aspect, the invention is directed to a system for secure sharing of artificial intelligence (Al) models, comprising one or more processors and memory storing instructions that, when executed by the one or more processors, cause the system to provide a secure clean room environment for sharing Al models between multiple parties, implement access control mechanisms comprising multi-factor authentication and role-based access control for governing access to the Al models within the clean room environment, create isolated sandbox environments within the clean room environment where authorized parties can test and use the Al models, monitor and log all activities performed within the clean room environment, enforce usage controls on the Al models, wherein the usage controls comprise at least one of time period limitations, execution frequency limitations, and use-case restrictions, and maintain end- to-end encryption of the Al models and associated data within the clean room environment.

[0022] In another aspect, the invention is directed to a method for secure sharing of artificial intelligence (Al) models, comprising the steps of providing a secure clean room environment for sharing Al models between multiple parties, implementing access control mechanisms comprising multi-factor authentication and role-based access control for governing access to the Al models within the clean room environment, creating a dynamic intent-based filteringAttorney Docket No. RAMP-00314-WOmechanism using multi-agent Al governance, creating isolated sandbox environments within the clean room environment where authorized parties can test and use the Al models, monitoring and logging all activities performed within the clean room environment, enforcing usage controls on the Al models, wherein the usage controls comprise at least one of time period limitations, execution frequency limitations, and use-case restrictions, enabling a bitmap compression and metadata separation system that enables structured access control by processing and segmenting large-scale datasets, and maintaining end-to-end encryption of the Al models and associated data within the clean room environment.

[0023] In another aspect, the invention is directed to a non-transitory computer-readable storage medium storing instructions that, when executed by one or more processors, cause the one or more processors to provide a secure clean room environment for sharing Al models between multiple parties, implement access control mechanisms comprising multi-factor authentication and role-based access control for governing access to the Al models within the clean room environment, create a dynamic intent-based filtering mechanism using multi-agent Al governance, create isolated sandbox environments within the clean room environment where authorized parties can test and use the Al models, monitor and log all activities performed within the clean room environment, enforce usage controls on the Al models, wherein the usage controls comprise at least one of: time period limitations, execution frequency limitations, and use-case restrictions, enable a bitmap compression and metadata separation system that enables structured access control by processing and segmenting large- scale datasets, and maintain end-to-end encryption of the Al models and associated data within the clean room environment.

[0024] These and other features, objects and advantages of the present invention will become better understood from a consideration of the following detailed description of the preferred embodiments and appended claims in conjunction with the drawings as described following:BRIEF DESCRIPTION OF DRAWINGS

[0025] Fig. 1 is an overall architectural diagram for an embodiment of the present invention.

[0026] Fig. 2 is a schematic for a computing component of a computing cluster for implementing an embodiment of the present invention.Attorney Docket No. RAMP-00314-WODETAILED DESCRIPTION OF THE INVENTION

[0027] Before the present invention is described in further detail, it should be understood that the invention is not limited to the particular embodiments described, and that the terms used in describing the particular embodiments are for the purpose of describing those particular embodiments only, and are not intended to be limiting, since the scope of the present invention will be limited only by the claims.

[0028] With reference now to Figure 1, an Al model sharing clean room 100 is provided within a cloud environment by solution provider 114. Through solution provider 114, access to clean room 100 is governed by a multi-layered authentication process 102, including multi-factor authentication (MFA), role-based access control (RBAC), and biometric verification. Partners are assigned specific roles with defined permissions, ensuring that they can only perform actions aligned with their responsibilities. Two partners 116 and 118 are shown in the illustrated embodiment, but the invention is not so limited.

[0029] All activities within the clean room are tracked and logged by monitoring module 104.Access logs show who accessed the clean room and when. Action logs show what actions were performed within the clean room. Usage analytics provide detailed reports on model usage and performance. These logs are stored securely and can be audited to ensure compliance with policies and agreements.

[0030] Data within the clean room is protected through robust encryption methods by encryption module 106. One method is encryption at rest, whereby Al models and data stored in the clean room are encrypted using industry-standard encryption algorithms. Another is encryption in transit, by which data transmitted between the clean room and external entities is encrypted using secure communication protocols.

[0031] Partners 116, 118 are provided with isolated environments where they can perform various actions in the form of sandbox environments 108. One action is the running of test Al models, whereby partners may run simulations and tests without affecting the original model. Another is the development of applications, such that partners may build and test applications that utilize the Al models within a secure and isolated environment. These sandboxed environments ensure that any changes or experiments do not compromise the security or integrity of the original Al models.

[0032] Clean room 100 includes tools to ensure adherence to data privacy regulations and contractual agreements. Policy enforcement by compliance management 112 providesAttorney Docket No. RAMP-00314-WOautomated enforcement of data usage policies. Compliance verification includes regular audits and checks to verify compliance with regulations such as GDPR, CCPA, and HIPAA. Reporting provides for generation of compliance reports for stakeholders and regulatory bodies.

[0033] Clean room 100 includes tools to ensure adherence to usage controls as required by the model owner in the form of usage controls 110. With time limit enforcement, the model may be made available for a certain period of time only. With execution limit enforcement, the model may be only allowed to be invoked a certain number of times per unit time or at a certain scale of instances, for example. With use-case enforcement, the model may be only allowed for certain use-cases or certain types of data sets or certain vertical industries. With named entity enforcement, the model may be only allowed to be used by certain named companies or disallowed for certain named competitors.

[0034] The control layer (the "guardrails") of the disclosed embodiment supports multi-tier authorization levels to define data accessibility. At one tier, full data authorizations applied for internal Al models under strict compliance. Filtered insights may be applied for external partners using the provider's Al exchange service, restrictive queries are used to block sensitive data exposure while allowing general Al responses.

[0035] Context-aware filtering applies the guardrails dynamically, based on user identity and role; query intent; data sensitivity classification; and regulatory compliance requirements (e.g., GDPR, CCPA, and HIPAA).

[0036] Bitmap compression and metadata separation provide for additional control. The system uses bitmap-based compression to efficiently process large datasets while maintaining structured access. Metadata is stored separately from primary data, enabling finer control over what Al models are able to retrieve. This also allows dynamic control over query granularity, limiting Al models to accessing only pre-approved dataset segments.

[0037] The system according to certain embodiments uses adaptive response management.This includes multi-agent governance to validate permissible Al-generated responses before output; redact, restructure, or restrict responses that violate data policies; and apply intentbased constraints using RAG. Adaptive response management may also include real-time logging and audit trails for compliance generation.

[0038] Secure clean room 100 for Al model sharing as described herein can be utilized in various scenarios. One scenario is collaborative research, enabling research institutions to share internal or external Al models without risking data privacy for raw data by automating policyAttorney Docket No. RAMP-00314-WOenforcement for Al-driven analytics with respect to enterprise data. Another scenario is enterprise partnerships, allowing businesses to share Al models with partners while ensuring compliance with usage agreements and corporate data policies. Another scenario is regulatory compliance, ensuring that Al models are used in accordance with data privacy regulations.

[0039] The invention in the disclosed embodiments offers several advantages over traditional model-sharing methods. One advantage is enhanced security, through which Al models are protected from unauthorized access and usage. Another advantage is the controlled environment, which ensures that models are used only for their intended purposes. Another advantage is compliance assurance, which helps organizations comply with data privacy regulations and contractual agreements. Another advantage is flexibility, by which isolated environments are provided for testing and development.

[0040] In the disclosed embodiments, the invention addresses the critical need for secure and controlled sharing of Al models. By providing a secure clean room environment, it ensures that Al models can be shared with partners without compromising security, privacy, or compliance. This solution fosters collaboration in Al development while protecting valuable intellectual property associated with the Al models.

[0041] In the disclosed embodiments, the invention enhances the compute environment via a controlled environment that enforces strict operational constraints, ensuring that Al models are only executed within predefined parameters and for their intended use cases in a computational manner. Flexibility is provided via containerized or virtualized environments that isolate development, testing, and production workflows. These isolated environments enable safe experimentation with new model architectures, hyperparameter tuning, and software updates without disrupting live operations. This modular approach ensures system integrity while facilitating rapid iteration and continuous integration / continuous deployment (CI / CD) pipelines. Together, these capabilities optimize computational resource allocation, reduce the attack surface, and streamline compliance processes, ultimately improving the overall efficiency, security, and reliability of the compute environment.

[0042] The methods described herein may in various embodiments be implemented by any combination of hardware and software. For example, in one embodiment, the methods may be implemented by a computer system or a collection of computer systems, each of which includes one or more hardware processors executing program instructions stored on a computer- readable physical storage medium coupled to the hardware processors. These computerAttorney Docket No. RAMP-00314-WOsystems may be part of a cloud computing networked environment. The program instructions may implement the functionality described herein (e.g., the functionality of various hardware servers and other components that implement the network-based cloud and non-cloud computing resources described herein). The various methods as illustrated in the figures and described herein represent example implementations. The order of any method may be changed, and various elements may be added, modified, or omitted.

[0043] Figure 2 is a block diagram of an example computer hardware system, according to various embodiments. Computer system 140 may implement a hardware portion of a cloud computing system as forming parts of the various implementations of the present invention. Computer system 140 may be any of various types of hardware devices, including, but not limited to, a commodity server, personal computer system, desktop computer, laptop or notebook computer, mainframe computer system, handheld computer, workstation, network computer, a consumer device, application server, physical storage device, telephone, mobile telephone, or in general any type of computing node, compute node, compute device, and / or hardware computing device.

[0044] Computer system 140 includes one or more hardware processors 140a, 141b...141n (any of which may include multiple processing cores, which may be single or multi-threaded) coupled to a physical system memory 142 via an input / output (I / O) interface 144. Computer system 140 further may include a network interface 146 coupled to I / O interface 144. In various embodiments, computer system 140 may be a single processor system including one hardware processor 141a, or a multiprocessor system including multiple hardware processors 141a, 141b...l41n.

[0045] Processors 141a, etc. may be any suitable processors capable of executing computing instructions. For example, in various embodiments, processors 141a, etc. may be general- purpose or embedded processors implementing any of a variety of instruction set architectures.

[0046] In multiprocessor systems, each of processors 141a, etc. may commonly, but not necessarily, implement the same instruction set. The computer system 140 also includes one or more hardware network communication devices (e.g., network interface 146) for communicating with other systems and / or components over a communications network, such as a local area network, wide area network, or the Internet. For example, a client application executing on system 140 may use network interface 146 to communicate with a server application executing on a single hardware server or on a cluster of hardware servers thatAttorney Docket No. RAMP-00314-WOimplement one or more of the components of the systems described herein in a cloud computing environment as implemented in various sub-systems. In another example, an instance of a server application executing on computer system 140 may use network interface 146 to communicate with other instances of an application that may be implemented on other computer systems.

[0047] In the illustrated embodiment, computer system 140 also includes one or more physical persistent storage devices 148 and / or one or more I / O devices 150. In various embodiments, persistent storage devices 148 may correspond to disk drives, tape drives, solid-state memory or drives, other mass storage devices, or any other persistent storage devices. Computer system 140 (or a distributed application or operating system operating thereon) may store instructions and / or data in persistent storage devices 148, as desired, and may retrieve the stored instructions and / or data as needed. For example, in some embodiments, computer system 140 may implement one or more nodes of a control plane or control system, and persistent storage 148 may include the solid-state drives (SSDs) attached to that server node. Multiple computer systems 140 may share the same persistent storage devices 148 or may share a pool of persistent storage devices, with the devices in the pool representing the same or different storage technologies, including such technologies as described above.

[0048] Computer system 140 includes one or more physical system memories 142 that may store code / instructions 143 and data 145 accessible by processor(s) 141a, etc. The system memories 142 may include multiple levels of memory and memory caches in a system designed to swap information in memories based on access speed, for example. The interleaving and swapping may extend to persistent storage devices 148 in a virtual memory implementation, where memory space is mapped onto the persistent storage devices 148. The technologies used to implement the system memories 142 may include, by way of example, static randomaccess memory (RAM), dynamic RAM, read-only memory (ROM), non-volatile memory, solid- state memory, or flash-type memory.

[0049] As with persistent storage devices 148, multiple computer systems 140 may share the same system memory systems 142 or may share a pool of system memories 142. System memory or memory systems 142 may contain program instructions 143 that are executable by processor(s) 141a, etc. to implement the routines described herein.

[0050] In various embodiments, program instructions 143 may be encoded in binary, Assembly language, any interpreted language such as Java, compiled languages such as C / C++, or in anyAttorney Docket No. RAMP-00314-WOcombination thereof; the particular languages given here are only examples. In some embodiments, program instructions 143 may implement multiple separate clients, server nodes, and / or other components.

[0051] In some implementations, program instructions 143 may include instructions executable to implement an operating system (not shown), which may be any of various operating systems, such as UNIX, LINUX, Solaris™, MacOS™, or Microsoft Windows™. Any or all of program instructions 143 may be provided as a computer program product, or software, that may include a non-transitory computer-readable storage medium having stored thereon instructions, which may be used to program a computer system (or other electronic devices) to perform a process according to various implementations. A non-transitory computer-readable storage medium may include any mechanism for storing information in a form (e.g., software or processing application) readable by a machine (e.g., a physical computer).

[0052] Generally speaking, a non-transitory computer-accessible medium may include computer-readable storage media or memory media such as magnetic or optical media, e.g., disk or DVD / CD-ROM, coupled to or in communication with computer system 140 via I / O interface 144. A non-transitory computer-readable storage medium may also include any volatile or non-volatile media such as RAM or ROM that may be included in some embodiments of computer system 140 as system memory 142 or another type of memory. In other implementations, program instructions may be communicated using optical, acoustical or other form of propagated signals (e.g., carrier waves, infrared signals, digital signals, etc.) conveyed via a communication medium such as a network and / or a wired or wireless link, such as may be implemented via network interface 606. Network interface 146 may be used to interface with other devices 142, which may include other computer systems or any type of external electronic device.

[0053] In some embodiments, system memory 142 may include data store 145, as described herein. In general, system memory 142 and persistent storage 148 may be accessible on other devices 142 through a network and may store data blocks, replicas of data blocks, metadata associated with data blocks, and / or their state, database configuration information, and / or any other information usable in implementing the routines described herein.

[0054] In one embodiment, I / O interface 144 may coordinate I / O traffic between processors 141a, etc., system memory 142, and any peripheral devices in the system, including through network interface 146 or other peripheral interfaces. In some embodiments, I / O interface 144Attorney Docket No. RAMP-00314-WOmay perform any necessary protocol, timing or other data transformations to convert data signals from one component (e.g., system memory 142) into a format suitable for use by another component (e.g., processors 141a, etc.).

[0055] In some embodiments, I / O interface 144 may include support for devices attached through various types of peripheral buses, such as a variant of the Peripheral Component Interconnect (PCI) bus standard or the Universal Serial Bus (USB) standard, as examples. Also, in some embodiments, some or all of the functionality of I / O interface 144, such as an interface to system memory 142, may be incorporated directly into processor(s) 141a, etc.

[0056] Network interface 146 may allow data to be exchanged between computer system 140 and other devices attached to a network, such as other computer systems (which may implement one or more storage system server nodes, primary nodes, read-only node nodes, and / or clients of the database systems described herein), for example. In addition, I / O interface 144 may allow communication between computer system 140 and various I / O devices 150 and / or remote storage 148. Input / output devices 150 may, in some embodiments, include one or more display terminals, keyboards, keypads, touchpads, scanning devices, voice or optical recognition devices, or any other devices suitable for entering or retrieving data by one or more computer systems 140. These may connect directly to a particular computer system 140 or generally connect to multiple computer systems 140 in a cloud computing environment, grid computing environment, or other system involving multiple computer systems 140.

[0057] Multiple input / output devices 150 may be present in communication with computer system 140 or may be distributed on various nodes of a distributed system that includes computer system 140. In some embodiments, similar input / output devices may be separate from computer system 140 and may interact with one or more nodes of a distributed system that includes computer system 140 through a wired or wireless connection, such as over network interface 146. Network interface 146 may commonly support one or more wireless networking protocols (e.g., Wi-Fi / I EEE 802.11, or another wireless networking standard).

[0058] Network interface 146 may support communication via any suitable wired or wireless general data networks, such as other types of Ethernet networks, for example. Additionally, network interface 146 may support communication via telecommunications / telephony networks such as analog voice networks or digital fiber communications networks, via storage area networks such as Fibre Channel SANs, or via any other suitable type of network and / or protocol. In various embodiments, computer system 140 may include more, fewer, or differentAttorney Docket No. RAMP-00314-WOcomponents than those illustrated (e.g., displays, video cards, audio cards, peripheral devices, or an Ethernet interface).

[0059] Any of the distributed system embodiments described herein, or any of their components, may be implemented as one or more network-based services in the cloud computing environment. For example, a read-write node and / or read-only nodes within the database tier of a hardware database system may present database services and / or other types of physical data storage services that employ the distributed storage systems described herein to clients as network-based services.

[0060] In some embodiments, a network-based service may be implemented by a software and / or hardware system designed to support interoperable machine-to-machine interaction over a network. A web service may have an interface described in a machine-processable format. Other systems may interact with the network-based service in a manner prescribed by the description of the network-based service's interface. For example, the network-based service may define various operations that other systems may invoke, and may define a particular application programming interface (API) to which other systems may be expected to conform when requesting the various operations.

[0061] In various embodiments, a network-based service may be requested or invoked through the use of a message that includes parameters and / or data associated with the network-based services request. Such a message may be formatted according to a particular markup language such as Extensible Markup Language (XML), and / or may be encapsulated using a protocol. To perform a network-based services request, a network-based services client may assemble a message including the request and convey the message to an addressable endpoint (e.g., a Uniform Resource Locator (URL)) corresponding to the web service, using an Internet-based application layer transfer protocol such as Hypertext Transfer Protocol (HTTP).

[0062] Unless otherwise stated, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0063] Although any methods and materials similar or equivalent to those described herein can also be used in the practice or testing of the present invention, a limited number of the exemplary methods and materials are described herein. It will be apparent to those skilled in the art that many more modifications are possible without departing from the inventive concepts herein.Attorney Docket No. RAMP-00314-WO

[0064] All terms used herein should be interpreted in the broadest possible manner consistent with the context.

[0065] When a grouping is used herein, all individual members of the group and all combinations and sub-combinations possible of the group are intended to be individually included.

[0066] When a range is stated herein, the range is intended to include all sub-ranges within the range, as well as all individual points within the range.

[0067] When "about," "approximately," or like terms are used herein, they are intended to include amounts, measurements, or the like that do not depart significantly from the expressly stated amount, measurement, or the like, such that the stated purpose of the apparatus or process is not lost.

[0068] All references cited herein are hereby incorporated by reference to the extent that there is no inconsistency with the disclosure of this specification.

[0069] The present invention has been described with reference to certain preferred and alternative embodiments that are intended to be exemplary only and not limiting to the full scope of the present invention, as set forth in the appended claims.

Claims

1. Attorney Docket No. RAMP-00314-WO2.Claims1. A system for secure sharing of artificial intelligence (Al) models, comprising:4.one or more processors; and5.memory storing instructions that, when executed by the one or more processors, cause the system to:6.provide a secure clean room environment for sharing Al models between multiple parties; implement access control mechanisms comprising multi-factor authentication and role-based access control for governing access to the Al models within the clean room environment;7.create isolated sandbox environments within the clean room environment where authorized parties can test and use the Al models;8.monitor and log all activities performed within the clean room environment;9.enforce usage controls on the Al models, wherein the usage controls comprise at least one of: time period limitations, execution frequency limitations, and use-case restrictions; and maintain end-to-end encryption of the Al models and associated data within the clean room environment.

2. The system of claim 1, wherein the instructions further cause the system to generate compliance reports based on the monitored activities to verify adherence to data privacy regulations.

3. The system of claim 1, wherein the sandbox environments enable authorized parties to: run simulations using the Al models without affecting the original models; andAttorney Docket No. RAMP-00314-WO develop applications that utilize the Al models within an isolated environment.

4. The system of claim 1, wherein implementing access control mechanisms further comprises: assigning specific roles to authorized parties; and13.defining permissions associated with each role to restrict actions that can be performed within the clean room environment.

5. The system of claim 1, wherein monitoring activities comprises:15.tracking access logs indicating who accessed the clean room environment and when; tracking action logs indicating what actions were performed within the clean room environment; and16.generating usage analytics regarding the Al models usage and performance.

6. The system of claim 1, wherein maintaining end-to-end encryption comprises:18.encrypting the Al models and the associated data stored at rest using industry-standard encryption algorithms; and19.encrypting the associated data transmitted between the clean room environment and external entities using secure communication protocols.

7. The system of claim 1, wherein enforcing usage controls comprises:21.restricting usage of the Al models to specified time periods;22.limiting the number of times the Al models can be invoked per unit time;23.restricting usage of the Al models to specific use-cases or industry verticals; and Attorney Docket No. RAMP-00314-WO restricting usage of the Al models to specified entities while preventing usage by specified competitors.

8. A method for secure sharing of artificial intelligence (Al) models, comprising:25.providing a secure clean room environment for sharing Al models between multiple parties; implementing access control mechanisms comprising multi-factor authentication and role-based access control for governing access to the Al models within the clean room environment;26.creating a dynamic intent-based filtering mechanism using multi-agent Al governance; creating isolated sandbox environments within the clean room environment where authorized parties can test and use the Al models;27.monitoring and logging all activities performed within the clean room environment; enforcing usage controls on the Al models, wherein the usage controls comprise at least one of: time period limitations, execution frequency limitations, and use-case restrictions;28.enabling a bitmap compression and metadata separation system that enables structured access control by processing and segmenting large-scale datasets; and29.maintaining end-to-end encryption of the Al models and associated data within the clean room environment.

9. The method of claim 8, further comprising generating compliance reports based on the monitored activities to verify adherence to data privacy regulations.

10. The method of claim 8, wherein the sandbox environments enable authorized parties to: run simulations using the Al models without affecting the original models; and32.develop applications that utilize the Al models within an isolated environment. Attorney Docket No. RAMP-00314-WO 11. The method of claim 8, wherein implementing access control mechanisms further comprises: assigning specific roles to authorized parties; and33.defining permissions associated with each role to restrict actions that can be performed within the clean room environment.

12. The method of claim 8, wherein monitoring activities comprises:35.tracking access logs indicating who accessed the clean room environment and when; tracking action logs indicating what actions were performed within the clean room environment; and36.generating usage analytics regarding the Al models usage and performance.

13. The method of claim 8, wherein maintaining end-to-end encryption comprises: encrypting the Al models and the associated data stored at rest using industry-standard encryption algorithms; and38.encrypting the associated data transmitted between the clean room environment and external entities using secure communication protocols.

14. The method of claim 8, wherein enforcing usage controls comprises:40.restricting usage of the Al models to specified time periods;41.limiting the number of times the Al models can be invoked per unit time;42.restricting usage of the Al models to specific use-cases or industry verticals; and43.restricting usage of the Al models to specified entities while preventing usage by specified competitors. Attorney Docket No. RAMP-00314-WO 15. A non-transitory computer-readable storage medium storing instructions that, when executed by one or more processors, cause the one or more processors to:44.provide a secure clean room environment for sharing Al models between multiple parties; implement access control mechanisms comprising multi-factor authentication and role-based access control for governing access to the Al models within the clean room environment;45.create a dynamic intent-based filtering mechanism using multi-agent Al governance; create isolated sandbox environments within the clean room environment where authorized parties can test and use the Al models;46.monitor and log all activities performed within the clean room environment;47.enforce usage controls on the Al models, wherein the usage controls comprise at least one of: time period limitations, execution frequency limitations, and use-case restrictions;48.enable a bitmap compression and metadata separation system that enables structured access control by processing and segmenting large-scale datasets; and49.maintain end-to-end encryption of the Al models and associated data within the clean room environment.

16. The non-transitory computer-readable storage medium of claim 15, wherein the instructions further cause the one or more processors to generate compliance reports based on the monitored activities to verify adherence to data privacy regulations.

17. The non-transitory computer-readable storage medium of claim 15, wherein the sandbox environments enable authorized parties to:52.run simulations using the Al models without affecting the original models; and Attorney Docket No. RAMP-00314-WO develop applications that utilize the Al models within an isolated environment.

18. The non-transitory computer-readable storage medium of claim 15, wherein implementing access control mechanisms further comprises:54.assigning specific roles to authorized parties; and55.defining permissions associated with each role to restrict actions that can be performed within the clean room environment.

19. The non-transitory computer-readable storage medium of claim 15, wherein monitoring activities comprises:57.tracking access logs indicating who accessed the clean room environment and when; tracking action logs indicating what actions were performed within the clean room environment; and58.generating usage analytics regarding the Al models usage and performance.

20. The non-transitory computer-readable storage medium of claim 15, wherein maintaining end- to-end encryption comprises:60.encrypting the Al models and the associated data stored at rest using industry-standard encryption algorithms; and61.encrypting the associated data transmitted between the clean room environment and external entities using secure communication protocols.

Images