870: BUILDING ENGINEERING TRUST FOR RELIABLE AEROSPACE SYSTEMS

 

(Through AI Assurance, Model-Based Engineering, Certification and Tamper-Evident Technologies)

 

Trust in engineering is not simply a matter of reputation or confidence. In aerospace and defence, it means having evidence that a system will perform as intended, that its behaviour can be verified, that changes are controlled and that its integrity is maintained throughout its operational life.

Aviation has always operated on this principle. Every aircraft, flight-control system and avionics component must meet demanding safety and reliability requirements before entering service. However, as aerospace systems become increasingly software-intensive, connected and autonomous, demonstrating that level of confidence is becoming more challenging.

Artificial intelligence (AI), complex electronic hardware, digital engineering and global supply chains are changing how aerospace systems are designed, manufactured and maintained. Traditional verification and certification processes remain essential, but they must evolve to address technologies whose behaviour can be difficult to predict or explain.

The challenge is therefore not just to build systems that work, but to ensure that they remain understandable, verifiable, secure and reliable throughout their service life.

Four areas are particularly important: controlled integration of AI, Model-Based Engineering (MBE) and Model-Based Definition (MBD), aerospace design assurance standards, and technologies that protect systems against unauthorised modification.

 

Wrappers to Manage AI Trust Gap

AI is finding increasing application in aerospace, including predictive maintenance, automated inspection, navigation, surveillance, target recognition and autonomous mission management. Machine learning is particularly useful when systems need to interpret large volumes of complex data that conventional rule-based software may struggle to handle. However, introducing AI into safety-critical applications creates a significant assurance challenge.

Traditional software is developed against defined requirements. Its logic can be examined, functions traced to requirements and implementation verified through structured testing. Machine learning operates differently. A trained neural network derives its behaviour from training data and learned parameters rather than explicitly programmed rules. Although its performance can be tested across a range of scenarios, it is difficult to demonstrate that it will behave correctly under every possible operating condition. Its performance may also change when sensor quality, environmental conditions or input data differ from those encountered during training and validation. A high confidence score does not necessarily mean that an output is correct or safe.

One practical way of managing this uncertainty is through a controlled software architecture, often referred to as an AI wrapper or harness. The wrapper manages the AI model’s inputs and outputs, controls access to external data and functions, checks results and ensures that the model operates within defined limits. It can also incorporate independent monitoring, fault detection and fallback mechanisms.

The wrapper does not automatically make an AI model safe or certifiable. Its purpose is to place the model within an architecture where its limitations can be monitored and its influence controlled. This hybrid approach allows AI to perform tasks where machine learning offers clear advantages, while critical control functions and protective mechanisms remain within established engineering and verification processes. The wrapper, monitoring logic and interfaces must themselves be verified as part of the complete system.

 

Establishing a Reliable Digital Foundation with Model-Based Engineering/Definition (MBE/MBD)

Engineering trust also depends on consistency between what is specified, designed, manufactured and eventually maintained. In traditional engineering programmes, requirements, drawings, specifications and manufacturing information are often maintained in separate documents and systems. Although these remain useful, managing large volumes of information creates the risk of inconsistencies. A design change may be reflected in one document but missed in another, leaving manufacturing or maintenance teams working with outdated information.

Model-Based Engineering addresses this problem by using connected digital models throughout the product lifecycle. These models can represent requirements, system architecture, functional behaviour, physical design, manufacturing processes and verification activities. Linking them allows engineers to trace decisions and assess the impact of changes across the programme.

Model-Based Definition takes the approach further by making an annotated three-dimensional model the primary source of product definition. Instead of relying mainly on traditional two-dimensional drawings, the model contains Product Manufacturing Information (PMI), including dimensions, geometric dimensioning and tolerancing (GD&T), materials, surface finishes and manufacturing annotations. The objective is to establish an authoritative digital product definition that can be used by design, manufacturing, inspection and maintenance teams.

A Model-Based Enterprise extends this principle across the organisation and its supply chain. For aerospace programmes, where aircraft and components may remain in service for several decades, this continuity is particularly valuable. It improves configuration control, reduces interpretation errors and helps ensure that suppliers and maintenance organisations work from the same approved technical information.

MBE and MBD can also strengthen certification activities. Requirements can be linked to design features, simulations, verification results and approved configurations. When a change is proposed, its impact on the design and associated certification evidence can be assessed more systematically.

 

Established Standards for Aerospace Design Assurance

Aerospace is an industry in which reliability must be demonstrated through objective evidence rather than assumed. Certification authorities expect organisations to show that systems have been developed, verified and controlled in accordance with recognised safety and security requirements. Several standards form the basis of this approach.

DO-178C: Airborne Software Assurance.DO-178C, Software Considerations in Airborne Systems and Equipment Certification, is the principal industry standard for demonstrating the assurance of airborne software (its European equivalent is ED-12C). It defines objectives covering planning, software development, verification, configuration management, quality assurance and certification liaison. The level of rigour depends on the software’s contribution to aircraft safety. Design Assurance Levels range from Level A, associated with catastrophic failure conditions, to Level E, where software failure has no effect on aircraft safety. DO-178C is therefore much more than a coding standard. It establishes a disciplined process for demonstrating that airborne software meets its intended requirements and does not contain unintended functionality.

DO-331: Model-Based Development and Verification. DO-331 supplements DO-178C by providing guidance for model-based development and verification. Engineering tools are widely used to model system behaviour, simulate designs and, in some cases, generate software code. DO-331 recognises that models may form an important part of the development process and that verification must address the models themselves, their relationship with requirements, model coverage and generated software. This is particularly relevant to flight-control systems, guidance functions and other applications where simulation and executable models help engineers identify design problems before implementation.

DO-254: Airborne Electronic Hardware Assurance. Modern avionics rely on complex electronic hardware and programmable devices. DO-254, Design Assurance Guidance for Airborne Electronic Hardware, provides a framework for demonstrating the assurance of such hardware. Its processes cover planning, requirements capture, design, implementation, verification, configuration management and process assurance. The objective is to establish that electronic hardware performs its intended functions and that sufficient evidence exists to support its use in an aircraft. The FAA’s AC 20-152A recognises DO-254/ED-80 as an acceptable means of compliance for applicable airborne electronic hardware and provides additional guidance on areas such as custom devices, commercial off-the-shelf components and intellectual property. This is particularly important as avionics hardware becomes more sophisticated. A validated software algorithm cannot provide overall system assurance if the processing hardware, memory, interfaces and supporting electronics have not also been adequately verified.

DO-326A: Airworthiness Security. As aircraft become more connected to onboard networks, ground systems, maintenance equipment and external communication links, cybersecurity becomes an important part of airworthiness. DO-326A, Airworthiness Security Process Specification, together with its European counterpart ED-202A, addresses security threats that could affect aircraft safety. It provides a structured approach to identifying threats, assessing their potential impact and defining suitable security measures. These include access controls, network segregation, secure software updates, vulnerability management and protection against unauthorised modification.

Security must be considered during the design of the system rather than as an afterthought added on during development. In the case of connected aircraft and autonomous platforms, getting it right at the design stage is the only way to avoid the very real risk of a vulnerability undermining the safety of the system. The standards together provide a set of recognised processes for the assurance of software, model-based development, hardware and security. They don’t eliminate engineering risk, they just help the organisation to manage it and to demonstrate that it has done so via documented evidence.

 

Maintaining Integrity Through Tamper-Evident Technologies

Aerospace components pass through manufacturers, suppliers, maintenance organisations and operational units. During this process, hardware, firmware, software and configuration data may be exposed to accidental damage, unauthorised access or deliberate interference. Tamper-evident technologies help detect or record unauthorised physical and digital changes. Physical measures include numbered security seals, tamper-indicating labels, specialised coatings and markings that reveal whether an enclosure or component has been opened or altered. More advanced approaches use embedded sensing meshes, conductive circuits and intrusion-detection mechanisms.

Digital protection is just as important. Secure boot ensures that the device is booting from authorised software. Cryptographic signatures are used to authenticate firmware and software updates. Hardware roots of trust provide a secure base from which to do identity and integrity checks, with cryptographic hashes and signed audit records being used to detect changes to the configuration and data. All of this is particularly pertinent to AI enabled systems. An ML model which has been extensively validated is not going to be considered trustworthy if its parameters or software package are subsequently changed without authorisation.

In defence applications, anti-tamper technologies may also protect sensitive designs and critical programme technologies through secure key storage, tamper-sensing enclosures and controlled access to protected information. Such measures must be designed carefully. For example, automatically shutting down a system or erasing data following a detected intrusion must not introduce additional safety risks.

Tamper-evidence provides protection in conjunction with certification. Certification proves that the configuration in question satisfies its requirements. Integrity protection is what enables you to establish that the configuration has not been altered at a later date without detection.

 

Combining All the Elements

The real value of these disciplines comes from using them together rather than treating them as separate engineering initiatives. MBE and MBD provide a consistent digital definition of the system and establish traceability between requirements, design, manufacturing and verification.DO-178C, DO-331 and DO-254 provide established assurance processes for software, model-based development and complex electronic hardware. DO-326A addresses security risks that could affect airworthiness. AI wrappers, deterministic supervisory functions, runtime monitoring and fault-management mechanisms help control system behaviour during operation. Secure boot, cryptographic protection and tamper-evident technologies help preserve the integrity of the approved configuration. Each layer addresses a different source of risk. None is sufficient on its own.

 

Implications for India’s Aerospace and Defence Industry

For India’s aerospace and defence sector, engineering trust is becoming increasingly important as the country develops indigenous aircraft, UAVs, advanced avionics, space systems and autonomous platforms. Developing the technology is only part of the challenge. Indian organisations must also demonstrate that these systems are safe, secure, maintainable and suitable for operational use.

Wider adoption of MBE and MBD can improve engineering consistency, supplier integration and lifecycle data management. Building expertise in DO-178C, DO-331 and DO-254 will strengthen the ability of Indian organisations to develop and certify safety-critical avionics and electronic systems. Similarly, incorporating DO-326A principles and appropriate security controls into system architectures will help protect connected platforms against cyber threats and unauthorised modification. AI offers significant opportunities in perception, inspection, predictive maintenance and decision support. However, operational deployment must be supported by disciplined data management, model validation, uncertainty assessment and independent safety controls.

A hybrid architecture is a sensible way forward.  There is also an important question of technological sovereignty. Dependence on externally developed black-box models, proprietary tools and components with limited technical transparency can create difficulties in assurance, maintenance and long-term support.

India therefore needs to build domestic capability in model validation, certification, secure hardware, digital engineering and mission software. The objective should not be to move away from international standards, but to develop the expertise needed to meet them while retaining control over critical technologies, engineering data and mission-specific software. This will strengthen operational capability and improve the international credibility of Indian aerospace products.

 

Concluding Thoughts

As aerospace systems become more intelligent, connected and autonomous, demonstrating their reliability will require greater engineering discipline.

AI offers capabilities that conventional software cannot always provide, but its use must be supported by appropriate controls and independent verification. MBE and MBD improve the consistency and traceability of engineering information. Aerospace standards provide established methods for demonstrating software, hardware and security assurance, while tamper-evident technologies help maintain system integrity throughout its service life. These disciplines are intimately linked. Digital models are of no use unless their information is properly governed. Certification is of any value at all if the approved configuration is maintained. AI can only be trusted in safety-critical applications if you understand its limitations and its failure modes.

Ultimately, engineering trust is not about claiming that a system is reliable, intelligent or secure. It is about demonstrating how it works, proving that it meets its requirements, controlling how it changes and ensuring that its integrity is maintained in service. That principle has guided aviation for decades and will remain just as important in future.

 

Please Add Value to the write-up with your views on the subject.

 

2192
Default rating

Please give a thumbs up if you  like The Post?

 

For regular updates, please register your email here:-

Subscribe

 

 

References and credits

To all the online sites and channels.

Pics Courtesy: Internet

Disclaimer:

Information and data included in the blog are for educational & non-commercial purposes only and have been carefully adapted, excerpted, or edited from reliable and accurate sources. All copyrighted material belongs to the respective owners and is provided only for wider dissemination.

 

References: –

  1. National Institute of Standards and Technology. (2023). Artificial intelligence risk management framework (AI RMF 1.0) (NIST AI 100-1). U.S. Department of Commerce. https://doi.org/10.6028/NIST.AI.100-1
  2. Federal Aviation Administration. (n.d.). Technical discipline: Artificial intelligence – Machine learning. https://www.faa.gov/aircraft/air_cert/step/disciplines/artificial_intelligence
  3. National Institute of Standards and Technology. (2024). Artificial intelligence risk management framework: Generative artificial intelligence profile (NIST AI 600-1). U.S. Department of Commerce. https://doi.org/10.6028/NIST.AI.600-1
  4. European Union Aviation Safety Agency. (2024). Artificial intelligence concept paper: Issue 2. EASA.
  5. European Union Aviation Safety Agency. (2020). Concepts of design assurance for neural networks (CoDANN). EASA.
  6. International Organization for Standardization. (2023). Systems and software engineering—System life cycle processes (ISO/IEC/IEEE Standard No. 15288:2023). ISO.
  7. International Organization for Standardization. (2025). Industrial automation systems and integration—Product data representation and exchange—Part 242: Application protocol: Managed model-based 3D engineering (ISO Standard No. 10303-242:2025). ISO. https://www.iso.org/standard/84300.html
  8. American Society of Mechanical Engineers. (2026). Digital product definition data practices (ASME Standard No. Y14.41-2026). ASME.
  9. International Organization for Standardization. (2019). Industrial automation systems and integration—Product data representation and exchange—Part 239: Application protocol: Product life cycle support (ISO Standard No. 10303-239:2019). ISO.
  10. International Council on Systems Engineering. (2023). Systems engineering handbook: A guide for system life cycle processes and activities (5th ed.). Wiley.
  11. RTCA, Inc. (2011). Software considerations in airborne systems and equipment certification (RTCA DO-178C). RTCA.
  12. RTCA, Inc. (2011). Model-based development and verification supplement to DO-178C and DO-278A (RTCA DO-331). RTCA.
  13. RTCA, Inc. (2000). Design assurance guidance for airborne electronic hardware (RTCA DO-254). RTCA.
  14. RTCA, Inc. (2014). Airworthiness security process specification (RTCA DO-326A). RTCA.
  15. RTCA, Inc. (2018). Airworthiness security methods and considerations (RTCA DO-356A). RTCA.
  16. Federal Aviation Administration. (2017). Airborne software development assurance using EUROCAE ED-12( ) and RTCA DO-178( ) (Advisory Circular No. AC 20-115D). U.S. Department of Transportation. https://www.faa.gov/regulations_policies/advisory_circulars/index.cfm/go/document.information/documentID/1032046
  17. Federal Aviation Administration. (2022). Development assurance for airborne electronic hardware (Advisory Circular No. AC 20-152A). U.S. Department of Transportation. https://www.faa.gov/regulations_policies/advisory_circulars/index.cfm/go/document.information/documentID/1041323
  18. Federal Aviation Administration. (2011). Development of aircraft and systems (Advisory Circular No. AC 20-174). U.S. Department of Transportation.

849: ENGINEERING TRUST IN THE DIGITAL AEROSPACE ERA

 

Shared my views on the subject at the DSCI Summit

on 09 Sep 26

 

 

Aerospace is undergoing the deepest transformation since the shift from piston to jet propulsion, except that this one is digital rather than mechanical. Aerospace systems are becoming software-defined, networked, and AI-enabled.

Flight control laws, mission systems, maintenance procedures, and command networks are increasingly mediated by software, sensors, and, increasingly, machine learning models.  These models can make decisions or influence them at a speed no human operator can independently verify in real time.

In an era where aerospace systems are defined as much by lines of code as by laws of aerodynamics. A compromise of software, data or communications can have consequences comparable to physical damage. In this environment, the “Engineering Trust” is taking on a new connotation.

 

Transformation of the Definition of “Engineering Trust”.

Engineering trust means designing, testing, and building systems so predictably safe and transparent that humans and regulators can rely on them without hesitation. The digital aerospace era has transformed the scope of “engineering trust”. Beyond the stated definition, it now also includes “having complete confidence that digital tools, data networks, and artificial intelligence (AI) systems are safe, secure, and fully verified”.

Trust must therefore extend from the physical platform to code, data, identity, networks, AI and the supply chain. It is no longer merely a peripheral technical concern. Engineering trust has become a central focus. Engineering digital trust requires validating each link, including hardware, flight software, network communications, and operational execution. It must be demonstrated, verified, and maintained throughout the system lifecycle.

The challenges relate to both the software code and the Data.

    • Code-related matters emphasise the importance of trust, Verification, authentic updates, and isolating compromised software to enhance resilience.
    • Data-related matters include data poisoning, Model manipulation, Adversarial inputs, AI Hallucination, unreliable outputs, and supply-chain digital compromises.

 

Core Pillars of Digital Aerospace Trust

Safety, Reliability, and Airworthiness. Safety, reliability, and airworthiness require that digital systems (such as flight controls, mission systems, avionics, MRO platforms, and autonomous functions) act predictably in normal, failed, and off-nominal situations without compromising platform or mission safety.

Cybersecurity and Cyber Resilience. The term cybersecurity, with an expanding scope, has been transformed into “digital mission assurance”. Aircraft, satellites, ground systems, and supporting infrastructure must be protected against cyber threats. Protective systems must prevent, detect, contain, withstand, and recover from attacks. This protection is required across the product lifecycle and supply chain.

Software Assurance and Verification. The development, verification, and certification of flight-critical software must be thorough and should follow the certification procedures adopted by aviation regulatory bodies such as the FAA and EASA. Formal methods and mathematical verification can complement testing and provide greater assurance of correctness, especially for functions that require a high level of assurance. Assurance must also extend to software updates, digital twins, and the growing number of adaptive systems.

Data Integrity and Sovereignty. Ensuring data integrity and sovereignty means that navigation, telemetry, sensor, maintenance, and operational data must be accurate, genuine, and readily available. It must also be protected against manipulation or spoofing.  Solid data pipelines, anti-spoofing methods, and sensor integration (for example, combining GNSS, INS, and optical tracking) can increase resilience. Regarding data sovereignty, it is also necessary to control where the data is stored, who can access it, and which legal or regulatory system applies, especially for multinational and classified programs.

AI Trustworthiness and Bounded Autonomy. AI/ML is used in predictive maintenance, decision support, pilot assistance, and autonomous systems. In these cases, evidence of robustness, transparency, accountability, and suitable human oversight is required. Autonomous features must operate within clearly defined and verifiable safety limits; deterministic safety mechanisms or “wrappers” (protective software boundaries or guardrails that surround an adaptive AI algorithm to restrict its behaviour and ensure safety) should keep adaptive algorithms within their certified operational envelopes.

Human–Machine Trust. The trust humans place in automation requires that pilots, engineers, and operators have enough insight into the system’s status, limitations, uncertainty, and decision-making logic to know when to rely on the automation and when to intervene or take over.

Supply-Chain and Firmware Integrity. Trust must extend throughout the entire manufacturing process, from component sourcing to software development, integration, deployment, and maintenance.

Continuous Lifecycle Assurance.  Trust cannot be established once and then assumed; it must be maintained through configuration control, monitoring, vulnerability management, software updates, supplier changes, operational data, and evidence provided throughout the entire system lifecycle.

 

Policy Framework for Engineering Trust in the Digital Era

Appropriate measures are required to ensure that India’s digital ecosystem for the aerospace and defence sector is secure, sovereign, traceable, certifiable and internationally trusted. The framework should contribute to national security, speed up the adoption of digital engineering and AI, improve supply chain resilience, support certification and exports, and, at the same time, increase confidence in India’s expanding indigenous aerospace and defence industrial base.

Make Digital Engineering Safe and Trustworthy. Set up a reliable digital engineering framework for the aerospace and defence sectors to guarantee the safety, reliability, security, and integrity of digital models, software, hardware, and digital twins. Adopt internationally recognised aerospace standards where appropriate and create end-to-end digital traceability covering the entire process from requirements, through design, implementation, verification, to certification. At the same time, maximise the automation of verification and testing and incorporate cybersecurity throughout the entire engineering lifecycle.

Protect Manufacturing and MRO from Cyber Attacks. Build robust and cyber-secure manufacturing, maintenance, repair and overhaul (MRO), testing, and operating environments using zero-trust principles. Ensure strong identity management, multi-factor authentication, least-privilege access, network segmentation, continuous monitoring, and strict third-party controls are in place. Cybersecurity requirements must cover the entire supply chain, including measures to ensure production and maintenance can continue during a cyber incident.

Make AI Safe and Trustworthy. We should establish a governance and assurance framework so that AI may be safely introduced into the aviation and defence sectors. Clear boundaries must be set regarding the level of autonomy of AI, and there must be adequate human supervision for any decisions which are of safety or mission importance. AI systems must be thoroughly tested under normal, abnormal, and failure conditions, monitored throughout their operational life, and supported by appropriate measures for accountability, data provenance, transparency, security, and auditability.

Create Digital Traceability for Parts and Products. Set up a complete digital traceability system for important aerospace and defence parts and components. Give each item a unique digital identity and keep reliable records relating to its origin, certification, configuration, inspections, repairs, modifications, and full lifecycle history. Apply tamper-evident technologies and digital product passports, as appropriate, in order to enhance authenticity, prevent the use of counterfeit components, and meet regulatory and customer assurance requirements.

Control Sensitive Data and use the Sovereign Cloud. Establish a national framework which covers the classification, protection, storage, processing, and controlled sharing of sensitive, classified, proprietary and export-controlled information. Sensitive aerospace and defence data should remain under the proper control of the nation and the organisation. For important programmes, use sovereign or controlled cloud environments, while allowing secure international cooperation without jeopardising sensitive data or intellectual property.

Develop Certifiable Digital Engineering Capability. Build national and industrial capabilities in the areas of model-based engineering, model-based systems engineering, and digital twins. Ensure digital engineering practices conform to international aerospace standards and automate verification where possible. Create definitive digital evidence to support airworthiness certification, obtain regulatory approval, and gain international customer acceptance.

Set up a Zero-Trust Defence Industrial Base. Introduce a zero-trust approach to cybersecurity in all defence organisations and within their industry ecosystem. Protect critical networks and systems by implementing robust identity management, using multi-factor authentication, applying segmentation, applying the principle of least privilege, and maintaining continuous monitoring. Make cybersecurity assurance a fundamental requirement of defence procurement and supplier management, backed by a common framework for managing cyber risk across the defence industrial base.

Develop sovereign AI and Cloud Capabilities. Create sovereign AI and cloud infrastructure for sensitive aerospace and defence applications, ensuring the nation maintains control over critical data, models, infrastructure, and intellectual property. Develop secure AI capabilities for areas including UAVs, predictive maintenance, mission support, simulation, and other defence applications, while allowing controlled collaboration with international partners.

Establish Governance for AI in Safety-Critical Systems. A governance protocol is required for using AI in safety-critical systems.  It should contain clearly defined sector-specific rules.  It should also include requirements for human supervision, intervention, testing, monitoring, accountability, and assurance throughout the system lifecycle. The governance framework must align with established principles of trustworthy AI, including safety, security, transparency, privacy, fairness, robustness, and accountability.

 

Concluding Thoughts

A digital platform, network, or system that you do not trust should not be considered a capability but rather a liability. The principle in question—that trust is essential—has not changed with digitalisation; in fact, digitalisation has made it harder and more urgent to engineer trust.

Trust in the digital aerospace era should not be treated as a secondary issue or as a compliance item to be dealt with after the project has been completed; instead, it must be built into the system from the very first line of code, the very first component, and the very first hour of operator training. It must be maintained throughout the system’s entire lifecycle.

Engineering trust in the digital era is therefore not merely about aerospace system reliability. It is about ensuring that their safety and security are continuously demonstrable, auditable, and verifiable.

 

Please Add Value to the write-up with your views on the subject.

 

2192
Default rating

Please give a thumbs up if you  like The Post?

 

For regular updates, please register your email here:-

Subscribe

 

 

References and credits

To all the online sites and channels.

Pics Courtesy: Internet

Disclaimer:

Information and data included in the blog are for educational & non-commercial purposes only and have been carefully adapted, excerpted, or edited from reliable and accurate sources. All copyrighted material belongs to the respective owners and is provided only for wider dissemination.

 

 

References: –

  1. SAE International. “Guidelines for development of civil aircraft and systems”, SAE Aerospace Recommended Practices ARP4754B, 2023.
  1. International Civil Aviation Organisation. “Cybersecurity action plan”, 2022.
  1. Rose, S., Borchert, O., Mitchell, S., & Connelly, S., “Zero trust architecture”. Special Publication 800-207, National Institute of Standards and Technology, 2020.
  1. Boyens, J., Smith, A., Bartol, N., Winkler, K., Holbrook, A., & Fallon, M. “Cybersecurity supply chain risk management practices for systems and organisations”. Special Publication 800-161 Rev. 1, Update 1, National Institute of Standards and Technology, 2024.
  1. National Institute of Standards and Technology. “Secure software development practices for generative AI and dual-use foundation models”, NIST AI 600-1, supplementary guidance, 2024.
  1. Tabassi, E., “Artificial intelligence risk management framework (AI RMF 1.0)” (NIST AI 100-1), National Institute of Standards and Technology, 2023.
  1. Autio, C., Schwartz, R., Dunietz, J., Jain, S., Stanley, M., Tabassi, E., Hall, P., & Roberts, K., “Artificial intelligence risk management framework: Generative artificial intelligence profile”, NIST AI 600-1, National Institute of Standards and Technology, 2024.
  1. Lee, J. D., & See, K. A. “Trust in automation: Designing for appropriate reliance”, Human Factors, 46 (1), 50–80, 2004.
  1. Indian Computer Emergency Response Team. “Blueprint for reducing exposure and defending against AI-assisted vulnerabilities exploitation in digital infrastructure”, 2026.
  1. Ministry of Defence, Government of India. *Security manual for licensed defence industries, 2025.

English हिंदी