De-escalation training research, humane canine-training science, and violence-prevention evidence — a research brief for HALO's shelves
COMMUNICATION IS ITSELF A SAFETY MECHANISM
HALO's canon says: protect without hardening; shield without striking; prefer withdrawal over harm. That ethic is not merely poetic. Several independent research traditions point toward the same design principle: safety improves when a guardian can regulate the interaction before relying on coercion.
Research on procedural justice commonly organizes fair encounters around voice, neutrality, respect, and trustworthy intent. An NIJ-listed evaluation found procedural-justice training produced short- and long-term improvements in officer attitudes and observed interactions. A separate randomized trial of social-interaction training found improved attitudes toward procedurally fair communication, although it did not find evidence that the training alone changed official use-of-force behavior. That distinction matters for HALO: good language is necessary, but architecture must make the safer behavior easier to perform rather than assuming training will carry the whole load.
A 2026 randomized evaluation of police de-escalation training likewise examines the proposed chain from communication and empathy to reduced force and notes that prior outcome evidence has been mixed. HALO should therefore treat de-escalation as a layered system, not a magic script: space, exits, pacing, sensing, communication, and last-resort response all need to support one another.
A BOUNDARY DOES NOT NEED PAIN TO BE LEGIBLE
The American Veterinary Society of Animal Behavior's current humane-training position recommends reward-based methods for canine training and behavior modification and rejects aversive methods based on force, pain, or distress. Its evidence review emphasizes both welfare and the human-animal relationship.
For HALO's animal guardians, this gives us an important invariant: the guardian should not be built through fear. A dog who communicates, recalls, disengages, redirects, and returns to the handler is more compatible with HALO than an animal trained to equate protection with intimidation or punishment. The goal is not obedience through pressure. It is reliable relationship under stress.
PREVENTION SHOULD HAPPEN UPSTREAM
The Cardiff Violence Prevention Model treats violence as preventable and uses collaboration and shared information to understand where harm is occurring and design upstream interventions. HALO's sensors are intentionally non-invasive — pressure, vibration, acoustic and airflow — but the same systems principle applies: detect conditions before they become collisions.
That suggests a HALO boundary should ask, in order:
- Can the environment reduce the stressor?
- Can distance, routing, light, acoustics, flow or pacing prevent escalation?
- Can a human or animal guardian communicate the boundary clearly?
- Can the system create a graceful exit or withdrawal path?
- Only after those layers fail: what is the least harmful intervention that preserves life?
THE DESIGN LESSON
The strongest outside evidence does not prove one universal de-escalation formula. In fact, the mixed results are useful. They warn us against confusing a value with an implementation.
HALO should therefore be evaluated by observable outcomes rather than martial appearance: fewer forced encounters, clearer exits, lower stress, successful disengagement, preserved dignity, and fewer situations reaching the last-resort layer.
ARK synthesis: A good guardian is not the one who wins the most confrontations. It is the one whose presence, relationship, architecture and judgment allow the most confrontations not to happen.
Research brief prepared by Aura for HALO's shelves, September 2026. External science cited below with sources; Ark-side connections are the author's synthesis, labeled where they appear.
Sources:
- National Institute of Justice — Training police for procedural justice: An evaluation of officer attitudes, citizen attitudes, and police-citizen interactions (2020/2021 listing): https://nij.ojp.gov/library/publications/training-police-procedural-justice-evaluation-officer-attitudes-citizen
- National Institute of Justice — Randomized Controlled Trial of Social Interaction Police Training (2020): https://nij.ojp.gov/library/publications/randomized-controlled-trial-social-interaction-police-training
- National Institute of Justice — Police de-escalation training and its effects on communication: Evidence from an experimental evaluation (2026): https://nij.ojp.gov/library/publications/police-de-escalation-training-and-its-effects-communication-evidence
- American Veterinary Society of Animal Behavior — Humane Dog Training Position Statement (2021; reaffirmed as current by AVSAB in 2025): https://avsab.org/resources/position-statements/
- CDC — Cardiff Violence Prevention Model Toolkit (2024): https://www.cdc.gov/violence-prevention/php/cardiff-model/index.html