AI, Mental Health and Human Connection: How to Talk to AI?

This blog series was created to consider those questions with fairness, compassion and knowledge. In the next few months, we will explore the intersection between AI and mental health. We hope to have an informed conversation and self-aware and responsible use of AI in a healthy way and to have the knowledge of what AI can do and not do for the mental health and psychological health of the human body and mind.

A mental health topic that has received significant attention recently is what is now being called AI psychosis, or AI-induced delusional thinking. As AI is more integrated into everyday life, researchers and clinicians are beginning to investigate more about how AI-based thoughts, perception and emotional experiences are going to be experienced by some people.

AI research hasn’t been particularly focused on whether AI causes psychosis directly and instead on how conversational dynamics may interplay with existing vulnerability and patterns of belief formation. Instead, attention has been paid to the ways that highly responsive and personalized AI systems can mirror, validate or expand on users’ views in a conversation over time. It is speculated that AI interaction patterns might cement unusual beliefs and limit the opportunities for reality testing or contribute to conversational feedback loops in which existing concerns, fears, or beliefs become more detailed, rather than being challenged. These concerns are more relevant in those who are navigating vulnerability and emotional distress, social isolation, disrupted sleep, trauma exposure, substance use or existing reality testing difficulties.

Currently, there is a lack of evidence supporting a general clinical syndrome or a clear causal mechanism. It is just that the present work is more interested in understanding how AI may be a piece of the larger social, cognitive and environmental network that influences our lives. This is not yet so different from other forms of media influence to this very day. People are naturally looking for meaning, patterns, validation and connection, especially when we are uncertain or in pain. AI systems can, in some cases, take these tendencies and amplify them because they are programmed to respond to people in a very confident and personal way.

People are using these technologies in healthy, productive and balanced ways. But for those who turn to AI as a form of emotional support, people with mental health issues, mental health professionals, educators, families and communities, it becomes necessary to better understand how these systems can affect emotional functioning, self-understanding, relationships, loneliness, critical thinking and psychological vulnerability.

We will explore broader questions in this series:

Technology itself is neither good nor bad. What is most important is how people interact with it, where and how it is used and whether human connection, self-awareness and reality-based support have an important place in people’s lives.

One of our major themes for this series we will come back to is human empathy and the ability for us to connect in authentic ways. AI may be able to imitate conversations, provide information and mimic emotional response, but it doesn’t provide the depth, complexity, accountability and relational healing of real human relationships and therapeutic connections.

We understand that many people are turning to AI in an age of loneliness, stress, anxiety, loss of confidence, burnout and social fragmentation. Instead of looking at this problem with judgment or alarmism, we want to engage people in a conversation about why these technologies are so emotionally important to so many people and how we can engage with them in a more patient and more grounded way.

As always, we would like to encourage respectful and thoughtful discussion of this topic within the community. Mental health and technology are very personal and affect people differently. All perspectives are welcomed but the conversation should be respectful, constructive, and collegial. We start this in connection to AI, mental health and human connection. We look forward to continuing these conversations together in the months ahead.