Dependable AI - Autonomous Systems

There is ongoing research on developing techniques for verifying and validating the safety of AI systems. This includes approaches such as model checking, formal verification, and testing, as well as the use of simulation and emulation. Further, there is wide spread interest on safety-aware AI. The aim of this aspect is to ensure that AI systems are designed from the ground up to prioritize safety and can adapt to changing safety requirements over time. This requires a multidisciplinary approach that combines expertise from computer science, engineering, and safety-critical domains.

• AI in functional safety

• AI with functional safety

• Hazards and Risks

• AI and ML development process

• ISO PAS 8800

• IEC ISO TR 5469

Bear in mind that AI in functional safety is a very new topic in industry and science. A common process has not yet been established. But we are pushing the technology approach from the beginning in 2014.