SLEEC@run.time
Abstract
SLEEC@run.time addresses the challenge of enforcing ethical behaviors at runtime in autonomous systems through a structured ethics assurance process. At the core of this process is a subsystem that rigorously operationalizes ethical rules, specifically social, legal, ethical, empathetic, and cultural (SLEEC) requirements. This subsystem enables the dynamic evaluation, adaptation, and enforcement of ethically compliant behavior within a formally defined runtime model. In addition, by leveraging its flexible runtime model, SLEEC@run.time accommodates changes such as the addition or removal of SLEEC rules, ensuring a robust and evolvable approach to ethical assurance in autonomous systems.
Ethics Assurance Process
The workflow is organized into four main phases. The first three phases: (1) rule elicitation, (2) formal rule specification and analysis, and (3) enforcement subsystem development take place at design time. The fourth phase: (4) enforcement subsystem operation runs at runtime.
These phases are executed sequentially; however, the SLEEC ruleset can be revisited and evolved at any point in the process. Such evolution may be triggered by signals or observations arising during any phase, prompting a reassessment of the normative framework.
When rules are reassessed, a refinement workflow propagates updates to both the design-time and runtime stages, ensuring the system remains aligned with evolving normative expectations.
Examples
Firefighter Scenario
The firefighter scenario consists of an uncrewed aerial vehicle (UAV) designed to support wildfire and
urban fire response operations. The UAV collaborates with human firefighters, bystanders, and remote
teleoperators to improve situational awareness and early intervention. Its primary tasks include
detecting potential fires, such as warehouse fires, using a thermal camera, localizing the fire with a
depth camera, and reporting the situation by transmitting video footage of the surveyed building to
teleoperators. When appropriate, the UAV can also deploy onboard water spraying to help contain the fire
until human firefighters arrive on site.
In addition to these functional capabilities, the firefighter UAV must address SLEEC considerations that
arise from interacting with people. For example,
activating an onboard loudspeaker alarm near bystanders may cause social disturbances, while
transmitting video footage can raise legal or ethical privacy concerns when individuals are present in
the monitored area.
Yaman, Sinem Getir, et al.
More examples coming soon...
BibTeX
Soon ...