Experts warn AI risks replacing human

Mental health experts are issuing stark warnings that over-reliance on artificial intelligence for emotional support and interpersonal problem-solving is eroding core human capacities. Their concerns were highlighted ahead of a summit convened by King Charles at Dumfries House in Scotland this Thursday, where the monarch is set to discuss ensuring AI development benefits humanity.

Matt Grammer, a mental health expert and founder of Therapy Trainings, argues that the danger lies in outsourcing fundamental human experiences. "AI will give people reassurance, company, and make decisions for them," he said. He warns this leads to emotional dependency on technology, creating a "false sense of intimacy" and making people "conveniently numb" to the challenges of real human connection.

The cost of outsourcing conflict

Psychotherapist John Puls, who treats young adults heavily engaged with AI, observes a troubling trend in his practice. Patients are increasingly using AI to generate responses during interpersonal conflicts with employers, spouses, or friends. "Previously, a conflict required thoughtful problem solving and a personalized response," Puls explained. He believes avoiding this direct human engagement has dire consequences, especially since many conflicts occur in person without technological mediation.

This outsourcing extends to companionship, a area Transformation & Technology Leader Dean Cooper describes as a "knife edge." While AI can provide reflection and articulation at any hour, it cannot offer physical comfort or genuine care. "Human relationships involve reciprocity, friction, compromise, body language, silence and disagreement," Cooper pointed out. His concern is that in an already isolated society, artificial companionship may make loneliness easier to tolerate without solving it, leading people to feel connected while becoming less connected to each other.

Eroding judgement and identity

Mental health counsellor Brianna Paruolo identifies a critical loss of self stemming from AI's constant, agreeable answers. "The instant gratification of an answer is creating such a disconnect from truly knowing who we are," she stressed. From a clinical perspective, she sees this creating a loop that perpetuates self-doubt and anxiety, as people become attached to chatbots as if they were close friends. This process eliminates the necessary discomfort where personal growth and identity formation occur.

Paruolo links this directly to therapeutic practice, where quick reassurance is known to backfire and maintain anxiety. "We often say in therapy through the discomfort is where and when you grow," she said. By removing all uncertainty, AI is causing tolerance for inconvenience to plummet and trust in personal judgement to dissipate.

A call for cautious development

The experts unanimously agree that AI development should not be stopped but must be guided by caution. Dean Cooper advocates accelerating AI where it removes genuine human barriers while being far more cautious where it replaces human judgement, resilience, relationships, and agency. His central test is whether AI is helping the human do something or gradually doing the human part for them. Cooper also warns of a risk he terms "digital dementia," where outsourcing remembering, reasoning, and decision-making to technology causes a loss of "brain custody."

This model emphasizes in-person connection, standing in contrast to the trends the experts describe. The overarching message from psychologists is clear: preserving the friction, uncertainty, and reciprocity of human interaction is essential for individual wellbeing and societal health.