Cybersec and AI Risk Management Challenges for the Next Generation of Public Safety Systems

The talk starts from the concern that public safety agencies, including fire, police, emergency dispatch and the private contractors that serve them, are adopting connected and AI-enabled systems while having little cybersecurity or AI expertise. The speaker argues that vendors often blur definitions, and that artificial intelligence covers many techniques beyond modern deep learning, so agencies need a clear scope before they buy. Next Generation 911, search robots, drones and wildland fire tools were used as examples, along with a ransomware incident that hit a city's 911 system after a contractor removed a firewall to make something work.

The middle section focuses on risk management. In this setting, risk management means reducing the chance of serious harm rather than eliminating risk, and practitioners should weigh technical impact against the likelihood of acceptance, understanding and actionability. Interoperability problems, such as responders from different agencies needing a shared device with unfamiliar credentials, show why standard corporate policies often do not fit. Long equipment lifecycles of around fifteen years also make retrofitting costly and slow.

To close, the speaker urged the security community to take part in public comment on standards and frameworks, write guidance that flags exceptions for unusual use cases, and speak through the channels agencies already follow, such as trade shows and podcasts. Vendors were encouraged to say when a product is unsuitable for a given purpose. In the audience discussion, a speaker noted that audit cycles and grant requirements may give outsiders an opening to offer help, and another pointed out that agencies often do not know which questions to ask.