You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Pilot for AI supported development process execution in safe context #805
Currently the S-CORE process_description defines workflows with templates, checklist, etc. as basis for the development approach used in S-CORE project. The current approach does not consider yet the support from AI based tools in safety context. This EPIC propose a concept how AI support could be incorporated in the current approach, but does still meet safety criteria. Further the EPIC proposes a pilot to implement the concept to present it at the next planned safety audit.
Processes itself are defined by a scoped and bounded workflow, triggered by defined inputs. They may use defined templates, checklists, guidelines, etc. to produce a defined output. The output is always a boundary, where a human actor has to be verify or validate the result and to approve it.
For the pilot implementation the AI supported Requirements Inspection as part of any Safety Requirements Workflow (Stakeholder, Feature, Component) in the process_description has been decided.
How
Implement e.g. a Requirements Inspection Agent to automate the inspection of any safety related requirements
Update process_description according the new approach (in addition, maybe in an experimental folder) to have the documentation in place
use ISO PAS 8800 and/or upcoming ISO/IEC CD TS 22440-1 for guidance (if available)
Present it add the next audit to get feedback for the generic concept
Estimates for realization
1 month
Category
Affects Detailed Design
Requirements / Architecture
Requirements / Architecture are not affected by this change
Activity