The recent shooting incidents in Philippine schools have understandably triggered alarm among parents, educators, and government officials. In August, a Grade 9 student at Ateneo de Zamboanga University shot and killed another student before taking his own life. The incident came only weeks after another fatal school shooting in Tacloban. These tragedies are complex, and we should be careful not to equate mental-health problems with violence. Most people experiencing mental-health difficulties are not violent. But these incidents should force us to ask whether we are doing enough to identify young people experiencing serious distress before they reach a crisis.

The immediate reaction has understandably been to tighten physical security. Schools have increased bag inspections, while proposals have surfaced for metal detectors and even transparent school bags. I understand why. When parents send their children to school, they expect them to come home safely. But imagine hundreds or thousands of students arriving at roughly the same time every morning and having every bag inspected for a gun, knife, or other prohibited object. Apart from creating long lines and requiring significant manpower, we are trying to detect the danger at the last possible moment — when the weapon has already reached the school gate. We are looking for the metal object instead of asking what happened to the child who decided to bring it.

Physical security will always be necessary, but it cannot be our only line of defense. Perhaps we should devote as much attention to identifying distress as we do to identifying weapons. The question then becomes whether technology, particularly artificial intelligence, can help us recognize warning signals earlier.

Every year, millions of Filipino employees undergo annual physical examinations. We measure blood pressure, cholesterol, and blood sugar. We take X-rays and check eyesight because we accept a fundamental principle of medicine: early detection allows early intervention. We do not wait for someone to suffer a heart attack before discovering hypertension. Yet our approach to mental health remains largely reactive. We often discover someone is struggling only when academic performance collapses, absenteeism increases, work deteriorates, or the individual finally asks for help. Sometimes the first obvious indication comes when a crisis has already occurred.

I believe it is time for Philippine schools and workplaces to explore periodic mental-health screening supported by AI. Students and employees could periodically and voluntarily complete short, confidential mental-health assessments using scientifically validated screening instruments. These could assess indicators associated with depression, anxiety, stress, burnout, loneliness, bullying, hopelessness, or suicide risk. This would not be radically different in principle from routinely measuring physical-health indicators. The objective is not to diagnose mental illness but to identify warning signals that deserve professional attention.

There are already established instruments demonstrating how simple initial screening can be. The US National Institute of Mental Health developed the Ask Suicide-Screening Questions or ASQ toolkit, consisting of four brief questions that can be administered in about 20 seconds. In its study, answering yes to one or more questions identified 97% of youth aged 10 to 21 who were at risk for suicide. A positive result does not constitute a psychiatric diagnosis. It signals the need for further assessment by a qualified person.

Artificial intelligence can potentially make this approach more powerful by looking not only at today’s answers but at changes over time. Imagine a student completing a confidential three-minute check-in every month. For several months everything appears normal, but gradually the answers change. Sleep deteriorates. Isolation increases. Academic pressure rises. Hopefulness declines. Eventually, responses indicate thoughts of self-harm. No single answer necessarily tells us the entire story, but the trajectory may tell us that someone needs attention.

A properly designed system could therefore work like an early-warning radar. Normal results could simply lead to preventive wellness resources. A significant deterioration could privately offer the student or employee an appointment with a guidance counselor, psychologist, employee assistance program, or other appropriate professional. Responses suggesting an immediate safety concern would activate an established human intervention protocol.

This is where we need to be very clear about AI’s role. It should not diagnose mental illness, predict who will commit violence, or decide who is suicidal. AI should detect patterns and help prioritize attention. Assessment, diagnosis, counseling, and intervention must remain in human hands.

This distinction is particularly important in the workplace, where mental-health concerns are also becoming increasingly significant. Employees face financial pressures, demanding workloads, family problems, long commutes, burnout, loneliness, and uncertainty about their careers. Ironically, AI itself is adding another source of anxiety as people wonder whether their jobs and skills will remain relevant. Philippine labor authorities have been emphasizing psychosocial safety and workplace mental-health programs, including confidentiality, referral, treatment, and employee support.

An AI-assisted system could help companies understand these problems earlier, but it must never become an employee-surveillance system. HR should not receive a list identifying employees as depressed, anxious, or potentially suicidal. Neither should companies secretly analyze employees’ private messages, e-mails, social-media activity, or personal devices in the name of wellness. An employee who admits feeling anxious must never wonder whether that answer will affect his promotion, performance evaluation, or job security.

The better model separates individual care from institutional intelligence. Personally identifiable alerts should be available only to appropriately authorized mental-health professionals operating under strict confidentiality and clearly defined escalation protocols. Management could instead receive anonymized, aggregated information. A CEO might discover, for example, that burnout indicators have increased significantly in one business unit over the past six months. That may reveal not an employee problem but a management problem involving excessive workloads, unrealistic targets, poor leadership, or an unhealthy culture.

Schools could similarly see broader trends without unnecessarily identifying individual students. Perhaps anxiety is increasing among graduating students. Perhaps loneliness is rising among freshmen. Perhaps indicators associated with bullying have suddenly increased in a particular year level. Administrators could respond with counseling, education, parent engagement, and other preventive interventions before the problem becomes more serious.

There are significant risks. AI systems can generate false positives and false negatives. Filipinos may express distress differently from populations on which international psychological instruments were developed. Mental-health data is also extremely sensitive. Any system would therefore require informed consent, data minimization, cybersecurity, limited retention, transparent rules governing access, clinical oversight, and human review. For children and adolescents, safeguards must be even stronger. The World Health Organization itself has warned that generative AI is already being used for mental-health support despite many systems not having been designed or clinically validated for that purpose.

The solution, therefore, is not an AI therapist. What I envision is much simpler: a Philippine mental-health early-warning system combining periodic validated screening, AI-assisted pattern detection, strict privacy protections, and guaranteed access to qualified human intervention. Government, schools, corporations, universities, technology companies, and mental-health organizations could collaborate on controlled pilots to determine whether such a model works in the Philippine context.

Bag inspections and metal detectors may prevent a weapon from entering a school on a particular morning. We should certainly improve physical security where necessary. But security at the gate deals with the final few minutes of a potentially much longer story. Mental-health intervention asks whether we could have reached the person weeks or months earlier.

We already understand this philosophy when it comes to physical health. We test blood pressure because we would rather discover the warning signs before the heart attack.

Perhaps it is time we applied the same thinking to the mind.

Dr. Donald Patrick Lim is the founding president of the Global AI Council Philippines and the Blockchain Council of the Philippines, and the founding chair of the Cybersecurity Council, whose mission is to advocate the right use of emerging technologies to propel business organizations forward. He is currently the president and COO of DITO CME Holdings Corp.