Today’s automotive and medical device innovators are increasingly integrating artificial intelligence (AI) and machine learning (ML) into safety-critical software systems. Because any failure in safety-critical systems can lead to death, serious injury, severe damage to equipment or property, environmental destruction, or other catastrophic events, applying and deploying AI- and ML-based components in these contexts can introduce significant risks.
As a result, companies involved in the development of safety-critical software systems face evolving challenges in several areas, such as:
- Integrating ML models into their software architecture.
- Defining requirements for highly complex yet inherently under-specified problems.
- Measuring the performance of AI/ML models.
- Performing quality assurance and certifying their AI- and ML-based components.
- Complying with the few published norms and standards that consider AI and ML in safety-critical systems.
- Adapting the development process in multiple locations and integrating roles — including data scientists, ML engineers, MLOps experts and data labelers — into the organization.
- Enabling effective communication and collaboration among data scientists, software engineers and the rest of the organization.
UL Solutions Software Intensive Systems services for AI/ML in safety-critical systems
At UL Solutions, we leverage our deep expertise in building safety-critical systems to offer a broad range of services that can help you navigate the challenges of integrating AI and ML in safety-critical systems in automotive, medical, aerospace, defense and transportation technologies.
Engineering support from UL Solutions Software Intensive Systems
Our expert engineers can help you define requirements in safety-critical systems’ inherently complex problems. We offer guidance in the technical aspects of where and how to integrate ML models into your software products. We can help you manage unexpected errors, which happen more often in AI-based components than in traditional software. And we can support you in deciding what frameworks to use or develop and what specialized hardware (such as graphic processing units (GPUs) or special AI accelerator chips) to use.
Through our AI/ML engineering support, we can also help you determine how to measure the performance of AI/ML models, execute automated testing over large amounts of data, create meaningful performance statistics, and perform quality assurance checks on data and labels.
UL Solutions Software Intensive Systems AI/ML consulting services
Because ML makes decisions based on data learning, which is fundamentally different from traditional software coded by humans, safety standards and norms for traditional safety-critical systems do not sufficiently address safety in ML-based systems. AI/ML applications are expanding and evolving rapidly, outpacing the release of applicable standards. We offer consulting services to help companies better understand the applicable norms and standards that have been published, which include:
- ISO 42001, Information Technology: Artificial Intelligence Management System, which specifies requirements for establishing, implementing, maintaining and improving the AI management system (AIMS) within organizations that develop or use AI-based products or services.
- ISO TR 5469, Artificial Intelligence: Functional Safety and AI Systems, which describes the properties, risk factors, methods and processes related to the use of AI inside a safety-related function, the use of non-AI safety-related functions to ensure safety for an AI-controlled equipment, and the use of AI systems to design and develop safety-related functions.
- ISO 21448, Road Vehicles: Safety of the Intended Functionality, which provides a framework and guidance on the design, verification and validation measures, as well as operation-phase activities, to achieve and maintain the safety of the intended functionality (SOTIF). (Learn more about UL Solutions Software Intensive Systems services relating to autonomy safety/SOTIF.)
Learn more about UL Solutions Software Intensive Systems services relating to autonomy safety/SOTIF.
Artificial intelligence trainings
Artificial intelligence trainings offered by UL Solutions Software Intensive Systems
From online trainings to hands-on workshops, we offer a variety of learning opportunities covering critical AI/ML and safety skills and topics.
Get connected with our team
Contact us to learn more about our AI/ML services, ask questions and get started.