Managing violence and aggression with emerging AI tools

AI's Push Against Rising Aggression

Workplace violence and aggression have escalated markedly in recent years, particularly in healthcare and education settings where frontline staff bear the brunt. In healthcare, workers are five times more likely to experience violence than those in other private-sector roles, a statistic that has hardened into a core operational and financial challenge rather than an occasional hazard.

The financial toll is staggering. Hospitals in the United States face roughly $18 billion each year from violence-linked expenses, covering staff turnover, workers' compensation claims, legal costs, and treatment for injured employees. This burden compounds existing pressures: workforce shortages, burnout, and high resignation intentions—with 61% of nurses surveyed in recent years indicating plans to leave their jobs within 12 months.

Parallel developments in education highlight similar strains, where challenging behaviours in schools threaten staff safety and require updated risk management approaches. The emergence of accessible AI tools, such as large language models for refining risk assessments and computer vision systems for monitoring behaviours, arrives at a moment when traditional manual methods fall short against increasing frequency and sophistication of incidents.

Deployments accelerated in 2025, with platforms from Teladoc Health analysing video and audio for threatening gestures, aggressive language, and facial cues in hospital environments, set for broader implementation in early 2026. Other systems repurpose existing surveillance cameras into proactive detectors of precursor actions like weapon draws or escalating confrontations, reducing reliance on post-incident review.

Yet the shift introduces trade-offs often glossed over. While AI promises fewer false positives through contextual analysis and quicker alerts, it raises questions about data privacy, potential biases in behavioural modelling, and the risk that over-dependence diminishes human judgment or critical thinking in high-stakes situations. Regulators and experts increasingly stress that effective use demands clear governance, human-in-the-loop protocols, and integration with broader safety cultures rather than standalone tech fixes.

Broader legislative and industry attention has intensified, with congressional discussions in early 2026 examining AI's role in transforming workplace safety from compliance-driven to predictive, even as voluntary frameworks dominate over mandatory rules in most jurisdictions.