Public Safety is vital for the functioning of societies: Without safety there is no freedom, no happiness, and no prosperity. The public good of safety matters to all of us, and therefore needs to be jointly shaped and maintained by all societal partners. Data generated by multiple agents play an increasingly important role in the prevention, preparedness and mitigation of harm or disaster. Multi-agency covers civil society (e.g., customers, citizen groups or NGOs), public bodies (e.g., municipalities, police forces, or first responders) and the private sector (e.g., security companies, IT companies or insurance providers). Our public safety is a multi-stakeholder phenomenon, characterized by an increasing amount of private-public partnerships, new alliances, and cross-sector collaborations.
With the arrival of Artificial Intelligence (AI), new opportunities arise for the beneficial use of algorithms trained from one or more of these multiple sources, for inclusive safety solutions which account for the different values, needs and goals of multiple agents. Yet, safeguarding human and public value is a condition for the development of such integral and crosscutting use of data: transparency, a certain level of mutuality in benefits, public accountability, respect for fundamental rights. Where this already presents a challenge in a purely public framework, the private-public agency mix increases the challenges, and AI will additionally affect the current public safety ecosystem.
The development of an ecosystem of trust regarding AI assisted public safety promotion is central to this ELSA Lab application. In a variety of use cases benefits and safeguards are analyzed against the private-public-machine agency backdrop. Based on a cutting-edge design-based approach, scalable and generalizable insights will be generated within our consortium and network representing the multi-agency in public safety. Outcome will be frameworks and products facilitating continuous multi-agency learning for the development of human-centered and ethically sound AI.
