Tamir Rice. John Crawford. Akai Gurley. Eric Garner. The list of men, especially men of color, who have been killed by police continues to grow with the recent death of Walter Scott. These tragedies are not isolated incidents. They are symptoms of a systemic problem: a police culture that trains and encourages officers to adopt a “warrior mindset.” As a former officer, I’ve been immersed in that culture. As a scholar who studies policing, I’ve seen how the warrior mindset, though adopted with the very best of intentions, has led to unnecessary violence and undermined police/community relations. In short, modern policing has developed a “warrior” problem.

Originally, the “warrior mindset” referred to the mental tenacity and attitude that officers, like soldiers, are taught to adopt in the face of a life-threatening struggle. In this context, the warrior mindset is a bone-deep commitment to survive a bad situation no matter the odds or difficulty. So narrowly defined, the concept is not problematic—indeed, it is commendable.

Unfortunately, the concept has mutated far beyond its original, limited meaning. Instead of applying to only the most dangerous and daunting situations, the warrior mindset now instructs officers on how to approach every aspect of their job. From their earliest days in the academy, would-be officers are told that their primary objective is to go home at the end of every shift. But, they are taught, they live in an intensely hostile world—one that is, quite literally, gunning for them. A recent article written by an officer for Police Magazine opens with this description: “The dangers we expose ourselves to every time we go [on duty] are almost immeasurable. We know this the day we sign up and the academy certainly does a good job of hammering the point home.” Training materials at the New Mexico Police Academy inform recruits that criminal suspects “are mentally prepared to react violently” adding, “[Y]ou could die today, tomorrow, or next Friday.”

In this worldview, officers are warriors combatting unknown and unpredictable—but highly lethal—enemies. They learn to be afraid. Officers don’t use that word, of course. Vigilant, attentive, cautious, alert, or observant are the terms that appear most often in police publications. But officers learn to be vigilant, attentive, cautious, alert, and observant because they are afraid, and they afraid because they’re taught to be.

As a result, officers learn to treat every individual they interact with as an armed threat and every situation as a deadly force encounter in the making. Every individual, every situation — no exceptions. A popular police training text offers this advice: “As you approach any situation, you want to be in the habit of looking for cover[] so you can react automatically to reach it should trouble erupt.” A more recent article puts it even more bluntly: “Remain humble and compassionate; be professional and courteous — and have a plan to kill everyone you meet.”

The warrior mindset creates a substantial, if invisible, barrier to good community policing. At its core, community policing relies on building “[c]ollaborative partnerships” so as to better identify community problems and “develop and evaluate effective responses.” To fulfill the promises of community policing, officers must establish meaningful relationships with individual community members. Imagine, however, that you are a rookie police officer driving down the street and looking to do some community policing. But you have been told (repeatedly) that your survival depends on believing that everyone you see — literally everyone — is capable of, and may very well be interested in, killing you. Would you actually get out of your car and approach someone? And if you did, would you stroll up to start a casual conversation or would you advance cautiously, ask for identification, run a criminal background check, and request consent to search . . . and then, maybe, try to start that casual conversation? The latter, of course, is what many officers are taught to do. My first ever “consensual encounter,” only hours into my first day of field training, followed exactly that pattern: after spending a few short minutes interacting with a young black transsexual, I left secure in the knowledge that she wasn’t armed and didn’t have any outstanding warrants. I knew almost nothing else about her undoubtedly complex and interesting life, and she knew even less about me. That approach, repeated over hundreds or thousands of police/civilian interactions, hinders the creation of meaningful, collaborative relationships.

Counterintuitively, the warrior mentality also makes policing less safe for both officers and civilians. Officers learn to both verbally and physically control the space they operate in. They learn that it is essential to set the proper tone for an encounter, and the tone that best preserves officer safety is widely thought to be one of “unquestioned command.” Even acting friendly, officers are told, can make them a target. But like the use of physical force, the assertive manner in which officers set the tone of encounter can also set the stage for a negative response or a violent interaction—one that was, from the start, avoidable. From the warrior perspective, the solution is simple: the people with whom officers interact must accede, respecting officers’ authority by doing what they are told. The failure to comply is confirmation that the individual is an enemy for the warrior to vanquish, physically if necessary. And remember that officers are trained to expect threats. Our brains are wired so that we see what we expect to see; given their training, it’s no surprise that officers react to threats that don’t actually exist. The result is avoidable violence.

Of course, violence is relatively uncommon; the Bureau of Justice Statistics reports police use physical force in about 1.5% of all encounters between police and public. But the small percentage masks a high absolute number: police interact with civilians about 63 million times a year. That means that, at a minimum, officers somewhere in the country use force close to a million times a year—or almost 2,600 times every day. Further, an officer who needlessly aggravates a situation doesn’t just increase the risk he faces in that specific encounter; he or she also increases the risk that other officers will face in future encounters. An aggressive approach in individual interactions can exacerbate underlying social tensions in a way that fuels a dangerous fire. This is not a new observation. As early as 1931, the Wickersham Commission, convened by President Herbert Hoover to investigate the problems with Prohibition enforcement, reported: “High-handed methods, shootings and killings, even where justified, alienate[] thoughtful citizens, believers in law and order. Unfortunate public expressions . . . approving killings and promiscuous shootings and lawless raids and seizures and deprecating the constitutional guarantees involved[] aggravate[] this effect.” In fact, of the ten most destructive and violent riots in United States history, fully half were responses to perceived police abuses. The expansive version of the warrior mentality promotes aggressive tactics; those tactics contribute to unnecessary violence; and the fierce rhetoric that follows further fans the flames.

The Warrior has created problems for law enforcement, but another model– the Guardian–may offer some solutions. So what’s the difference? Both Warriors and Guardians seek to protect the communities they serve, of course, but the guardian mindset takes both a broader and a longer view of how to achieve that goal. Put simply, the guardian mindset prioritizes service over crime-fighting, and it values the dynamics of short-term encounters as a way to create long-term relationships. It instructs officers that their interactions with community members must be more than legally justified; they must also be empowering, fair, respectful, and considerate. It emphasizes communication over command, cooperation over compliance, and legitimacy over authority. In the use-of-force context, the Guardian mindset emphasizes restraint over control, stability over action. But the concept is even broader; it seeks to protect civilians not just from crime and violence, but also from indignity and humiliation.

Here are some practical suggestions for training Guardian Officers instead of Police Warriors:

Non-enforcement Contacts. Officers should be required to initiate non-enforcement contacts with community members. Building on the “Good Strangers” and “Tact, Tactics and Trust” training that grew out of the Defense Advanced Research Projects Agency’s Strategic Social Interaction Modules training, a “non-enforcement contact” means an interaction in which officers are prohibited (except in emergencies, of course) from taking enforcement actions: no asking for identification, no running criminal history checks, no issuing tickets, and no making arrests. The purpose is threefold: give officers and community members the chance to get to know each other as individuals, emphasize the agencies’ visible commitment to community policing, and teach officers communications skills that they will use countless times over the course of their careers.

De-escalation Training. More than 98 per cent of police academies provide training in deadly force, less-lethal force, and less-than-lethal force. On average, recruits get over 160 hours of instruction in these topics. Fewer academies–88 per cent–train recruits in mediation or conflict resolution, and those offer, on average, a paltry 8 hours. For some officers—born talkers—that’s more than enough. Others find these skills harder to acquire. But talking to people, managing conflict without violence, is a learnable skill that all officers would benefit from having.

Integrated, Scenario-Based Training. Ask any police instructor for the ideal way to teach recruits physical skills and they’ll tell you that it is scenario-based role-playing training where they can practice in a realistic, responsive, and free-form environment. The same is true of de-escalation training. De-escalation is an effective tool in tense situations, but, like unarmed combat, it requires practice in a dynamic, high-pressure environment. Teaching de-escalation only in the classroom simply does not give officers confidence in their skills. Without that confidence, officers are far more likely to fall back on physical violence.

Informed Training. Right now, police training is highly responsive to a very limited set of situations: when an officer is seriously injured or killed, police trainers across the country engage in an in-depth analysis, in order to learn from the incident so it never happens again. We should demand the same approach when civilians are seriously injured or killed. Too many police agencies reflexively refuse to do anything that looks like second-guessing, in part because of concerns about legal liability and morale. That refusal is a missed opportunity to learn how to avoid future tragedy.

Tactical Restraint. The use of force, including deadly force, will sometimes be necessary. But when violence is avoidable and when avoiding it doesn’t sacrifice the police mission, officers should be required to use tactical restraint even when that means holding their position or temporarily withdrawing. From the guardian perspective, reducing the chances of finding a situation that requires violent response minimizes risk to police officers and the public. Officers are safer, which means that civilians are safer. I’ve described tactical restraint at more length in a recent piece in the Harvard Law Review Forum, so I’ll summarize by noting that tactical restraint encourages officers to work smarter, not harder, by relying more on good tactics and communication than on violence. This is why officers at most agencies are told to wait for back-up before physically making an arrest or tackling a running suspect. Using tactical restraint protects officers and civilians alike, which is exactly the approach Guardians should favor.

It will take more than the changes I’ve suggested here to change police culture and to heal the long-standing divide between law enforcement agencies and the communities they serve, particularly communities of color. Earning public trust will take years, if not decades, and it will require deep changes to police training, culture, and accountability mechanisms. But we can, we must, start now. We should start by rejecting the concept of the Police Warrior and adopting instead the Guardian Officer.

Our ideas can save democracy... But we need your help! Donate Now!

Seth Stoughton is a law professor at the University of South Carolina and a former police officer. A prior version of this article appeared in the Harvard Law Review Forum.