Is Artificial Intelligence Friend or Foe? It Depends

Core claim

AI is neither friend nor foe inherently; like other powerful forces in history it can promote or erode well-being depending on use. Its greatest psychological potential lies not in the answers it provides but in the questions it can ask — contemplative, open-ended questions that foster introspection, critical thinking, and resilience. Subtitle: “Don’t look to AI for the right answers, look to AI for the right questions.”

Framing the debate

• Two competing positions on how AI will change us psychologically and biologically:
  • It will rob us of our humanity, fostering passive consumption of and over-dependence on AI-generated information.
  • Like the Industrial Revolution, it will relieve us of burdens and has the potential to bring out the best in us — to make us more human.
• Everly’s answer to “friend or foe”: “It depends.”

The current state of AI

• AI does not originate thought or knowledge independently of retrieved data. It produces outputs through varying degrees of data recall and novel recombination of learned patterns.
• Even so-called novel outputs are built on previously acquired information, even when the specific output has never existed before — what Everly has termed “recombinant AI” in an earlier post (Psychology Today, 9 April 2026).
• The inherent weakness: AI cannot independently assess the validity of the data it retrieves or its own recombinant output, and so may provide misleading or erroneous information.

The risk of passive consumption

• From a public health perspective, Everly treats passive consumption as more problematic than erroneous output.
• The risks remain undetermined, but concerns include:
  • Inhibition of critical thinking
  • Functional dependence
  • Potential contributions to depression and isolation
  • Undesirable changes in brain morphology
• These outcomes have been associated with excessive social media use (Montag & Becker, 2023; Shensa et al., 2018), and the concern has generalized to AI.

The case for questions

Therapeutic alliance

• Even in its present form, AI appears capable of using active listening techniques to foster a therapeutic alliance comprising:
  • A sense of connection (being heard)
  • Empathy (being understood)
  • Cathartic ventilation
• Carl Rogers held these to be the necessary and sufficient conditions for self-improvement and therapeutic change. Possible net effects: conveyance of self-worth, optimism, and self-efficacy (Malouin-Lachance et al., 2025).

The Socratic precedent

• Roughly 2,000 years ago, Socrates pursued virtue and wisdom through challenging questions rather than answers, nurturing the most influential philosophers of his era.
• Consistent with the teachings of Confucius, he believed the highest form of wisdom arose from introspection and self-reflection — which questions can foster.

The neuroscience

• Open-ended questions recruit diverse neural networks across both hemispheres: the left hemisphere’s analytical reasoning linked to the right hemisphere’s integrative contextual awareness.
• This potentiates increased cognitive flexibility, emotional regulation, creativity, and resilience (Southwick & Charney, 2018).
• The most creative and resilient brains may be those engaging numerous and diverse integrated inter-hemispheric regions.
• The resilient brain is characterized by “elegant switching” between:
  • Default Mode Network activity (introspection and reflection)
  • Salience Network activity (hypervigilance and threat assessment)
  • (Iyer et al., 2026)
• Everly’s metaphor: contemplative questions are the gateway to this “symphony of creativity and resilience,” and AI seems capable of being its conductor.

Conclusion

• Over-reliance on AI to answer all our questions “has the potential to open Pandora’s Box.”
• Quoting Francis Bacon: “If a man will begin with certainties, he shall end in doubts; but if he will be content to begin with doubts, he shall end in certainties.”
• Questions let us look beyond ourselves and dream, but more importantly look inside ourselves and awaken abilities we may not have known we have.
• Closing line: “AI can help us be what AI itself cannot. Perhaps AI can help us be more ‘human.’”

Notes

• Published in the Psychology Today blog “When Disaster Strikes: Inside Disaster Psychology”; reviewed by Jessica Schrader.
• George S. Everly, Jr., Ph.D., ABPP, FACLP serves on the faculties of the Johns Hopkins Bloomberg School of Public Health and the Johns Hopkins School of Medicine.

References cited in the post

• Iyer, A., Seeley, S. H., Schreiber, Z., et al. (2026). Psychological Resilience in World Trade Center Responders is Associated with Effective Brain Network Connectivity. PsyArXiv. DOI: 10.31234/osf.io/4dbzx_v1
• Malouin-Lachance, A., Capolupo, J., Laplante, C., & Hudon, A. (2025). Does the digital therapeutic alliance exist? Integrative review. JMIR Mental Health, 12, e69294.
• Montag, C., & Becker, B. (2023). Neuroimaging the effects of smartphone (over-)use on brain function and structure. Psychoradiology, 3, kkad001.
• Shensa, A., et al. (2018). Social Media Use and Depression and Anxiety Symptoms: A Cluster Analysis. American Journal of Health Behavior, 42(2), 116–128.
• Southwick, S. M., & Charney, D. S. (2018). Resilience: The Science of Mastering Life’s Greatest Challenges. Cambridge University Press.