Artificial intelligence is emerging as a potential ally in addressing the mental health crisis, although experts warn about its risks and limitations.
In a world where, according to the WHO, anxiety and depression affect one in eight people and one in seven young people, mental health has become a global imperative. Against this backdrop, artificial intelligence (AI) has emerged as one of the most promising—and controversial—tools for tackling this silent crisis. Latin America is no exception to this tension: with fragmented healthcare systems and scarce resources, AI could either become an ally in democratizing access or a threat if not carefully regulated.
In recent years, generative AI tools—such as therapeutic chatbots—have proliferated. Far from being a marginal phenomenon, consultations related to emotional support are now among the leading reasons people use AI, alongside organizing their lives or finding purpose, according to a Harvard Business Review study (2025). This trend reflects a shift toward the emotional and existential dimensions of AI use, driven by advantages that are difficult to match: 24/7 availability, anonymity, low cost, and the possibility of speaking without fear of judgment.
But can an algorithm provide genuine emotional support? Are we ready to delegate our deepest emotions to a system trained on data?
AI applied to mental health does more than accelerate diagnoses. It also opens new avenues for preventive intervention, enables more personalized treatments, and can serve as a bridge for populations that are currently outside the healthcare system. Technologies such as immersive virtual reality, for example, are already being used to treat phobias, post-traumatic stress disorder, and social anxiety. And AI’s potential goes even further: it can help train professionals, create peer-support networks, and provide early warnings about critical symptoms.
However, enthusiasm coexists with serious concerns. The use of chatbots as “therapists” without supervision has raised alarms among organizations such as the American Psychological Association. In some cases, rather than challenging negative beliefs, these systems tend to reinforce them. And when vulnerable users—especially adolescents—are involved, the consequences can be severe.
At the clinical level, the limitations are evident: AI does not interpret nonverbal language, does not understand emotional context, and lacks ethical judgment. It cannot replace human empathy, provide therapeutic continuity, or offer deep cultural understanding. At best, it can serve as a complement, but never as a substitute for human connection.
In Latin America, the situation is paradoxical. On the one hand, AI could expand access in remote regions or areas with few professionals. On the other hand, only a handful of countries—such as Brazil, Chile, and Uruguay—have AI strategies that include healthcare, and in many cases implementation remains fragmented or in its early stages.
The broader context does not help. High levels of inequality, uneven connectivity, and low investment in public health mean that the risks of dehumanization, algorithmic bias, and technological exclusion are more pronounced than in other parts of the world. Without clear public policies and a robust regulatory framework, the use of AI in mental health could deepen—rather than reduce—existing inequalities.
Within this context, one of the main challenges is the use of data. For AI to be effective, it must be trained on representative information. Yet in Latin America, healthcare systems often lack digitized records, and existing data do not always reflect diversity in terms of gender, ethnicity, or socioeconomic conditions.
This creates an ethical dilemma: without local data, algorithms replicate models developed in the Global North, which can result in inaccurate diagnoses or inappropriate treatments. Privacy is another sensitive issue. Who controls emotional data? How accurate are algorithms at interpreting emotions? How can an adolescent who shares suicidal thoughts with a chatbot be protected?
In this scenario, multilateral organizations and governments have a fundamental role that goes far beyond financing. They must promote ethical standards, facilitate dialogue among different sectors, and ensure that digital transformation in healthcare is carried out with a human rights-based approach that embraces inclusion and cultural diversity. Promoting technology alone is not enough; it is essential to accompany it with regulations, talent development, and governance frameworks that protect what is most valuable—our mental health—in a time of rapid change.
Because if the future is going to be mediated by algorithms, then the question is not only what they can do for us, but also what values we want to guide their use. And that is a conversation that international organizations and governments can neither postpone nor observe from the sidelines.