Guessing who might be violent, then involving the police — what could go wrong?

Even as American cities are working to reduce the use of police in responding to mental health and social crises, the Biden administration is doubling down on an ineffective strategy that further entrenches law enforcement in these same spheres under the umbrella of violence prevention.

President Joe Biden's just-released National Strategy for Countering Domestic Terrorism features this approach in its plan to combat far-right violence. The Department of Homeland Security recently created a new Center for Prevention Programs and Partnerships, which will provide funds and support to local law enforcement, community groups and institutions such as universities to carry out such prevention efforts. Among its purposes is to identify people who may become violent and connect them with mental health and social services, often in cooperation with police.

The Homeland Security Department describes this as a "public health" approach, which may sound appealing. But decades of research show that we cannot reliably identify potentially violent people. And trying to do so will invite more police involvement in mental health and social services and bias against the same communities that bear the brunt of far-right violence, as a new report from the Brennan Center for Justice documents.

Many of the behaviors and traits the center identifies as markers of potential violence — being socially alienated, depressed, having a "grievance," for example — are both vague and common. Treating what are often adverse social conditions as potential police matters hurts efforts to support people struggling with these conditions.

The new center essentially puts a new label on Homeland Security's old Targeted Violence and Terrorism Prevention program, which Biden had promised to end. That program, in turn, was a rebrand of a war-on-terror-era program called Countering Violent Extremism, which broadly treated Muslim Americans as terrorism risks.

These earlier programs treated actions such as attending a mosque more frequently and being concerned about anti-Muslim discrimination or human rights abuses as reasons for criminal suspicion. While the Biden administration's disavowal of the heavy-handed targeting that marked the war on terror approach is welcome, the new program's prevention activities rest on the same flawed foundation and impose many of the same harms.

The Biden program claims its prevention model is evidence-based, but the very studies it cites say that predicting who will engage or attempt to engage in terrorism "is an unrealistic goal." Instead, government-run studies in this field claim to identify commonalities among those who have carried out violent attacks, labeling them risk factors and indicators that bear on whether a person is going to commit violence.

Nor does Homeland Security account for how race, religion and ethnicity influence who is tagged as dangerous. This holds true in schools, where discipline falls more heavily on children of color; in policing, where race often dictates who is targeted for enforcement; and in counterterrorism, where Muslims have borne the brunt of suspicion.

The new program is supposed to work with the Homeland Security Office of Civil Rights and Liberties to ensure rights are protected, but it has not specified any concrete safeguards. The program formally requires those receiving its grants to address privacy and civil rights concerns when applying for funds. But these requirements also existed in the earlier programs, without much effect.

Of course, people experiencing conditions that the program identifies as potential markers of violence could well benefit from mentorship programs or mental health treatment. But linking access to such services to a propensity for violent crime makes it less likely that people will seek out help when they need it.

A better path forward is to wall off security agencies such as Homeland Security from efforts to address the social problems the department frames as threats and leave these issues to people with the right expertise. One blueprint is the recently reintroduced Counseling Not Criminalization in Schools Act, which proposes funds to replace police in schools with social service providers such as teachers, counselors, social workers and nurses and prohibits the use of money for partnerships with law enforcement.

Homeland Security's Center for Prevention Programs and Partnerships is unlikely to help prevent violence by mixing health and social services in a law enforcement framework, but it will harm the communities it is trying to protect. The Biden administration should instead invest in badly needed social services through the agencies most equipped to provide them.