• Login
  • Register

Work for a Member organization and need a Member Portal account? Register here with your official email address.

Project

AI and Human Psychology

Copyright

Advancing Humans with AI (AHA)

Advancing Humans with AI (AHA)

Groups
​

The Advancing Humans with AI (AHA) research program at the MIT Media Lab explores research on AI and human psychology. 

We explore how AI shapes core dimensions of human experience, including motivation, loneliness, relationships, long-term thinking, learning, memory, reasoning, attention, and emotional wellbeing. Our work examines both the positive outcomes AI can foster (such as learning, agency, healthier relationships, and personal growth) and the negative outcomes it can produce (such as overreliance, cognitive debt, sycophancy, manipulation, false memories, loneliness, and emotional dependence). Rather than being uniformly optimistic or pessimistic, we run rigorous empirical studies and build new systems to understand how to design AI that supports human flourishing.

Critical Thinking, Reasoning, and Cognitive Debt

We study how AI can either augment human reasoning or quietly replace it. Our systems are designed to provoke reflection and questioning rather than deliver passive answers, and our empirical work measures the cognitive cost of overreliance on AI assistants and the durability of any gains in misinformation discernment.

Selected projects and publications:


Learning and Education

Our learning research goes beyond engagement to ask how AI can be designed to deepen understanding, support neurodivergent learners, adapt to individual cognitive states, and complement (rather than replace) human peers and teachers.

Selected projects and publications:

Loneliness, Relationships, and AI Companions

As people increasingly form emotional bonds with conversational AI, we study the consequences for loneliness, attachment, social connection, and the grief that arises when these AI relationships end. Our research examines when AI companionship supports wellbeing and when it risks deepening isolation, fostering unhealthy dependence, or producing real psychological harm.

Selected projects and publications:

Memory and Recollection

We study how AI can both augment and distort human memory. Our wearable systems help people recall details from real conversations, locate lost objects, and revisit lived experiences, while our empirical studies show how generative AI can also implant false memories and reshape how we remember the past.

Selected projects and publications:

Long-term Thinking and the Future Self

We design AI systems that help people connect with their future selves, take a longer time horizon, and make wiser decisions in the present. This work draws on psychological research on self-continuity, intertemporal choice, and prospective cognition.

Selected projects and publications:

Motivation and Engagement

We investigate how the design of AI systems shapes intrinsic motivation, curiosity, and engagement, especially in learning contexts. Our studies show that personalizing AI characters and contexts can meaningfully increase learner motivation, while poorly designed AI can erode it.

Selected projects and publications:

Beliefs, Trust, and the Placebo Effect of AI

People's prior beliefs about AI shape what they perceive AI to say and do. Our work shows that priming users' expectations can significantly change perceived trustworthiness, empathy, and effectiveness of the same underlying model, and that users often place misplaced trust in AI even in high-stakes settings such as medicine.

Selected projects and publications:

Wellbeing, Mental Health, and Psychological Risk

We are building benchmarks, frameworks, and simulation methods for evaluating AI not just by accuracy or efficiency, but by its impact on human emotional, social, and psychological wellbeing, including its potential to contribute to severe mental health harms.

Selected projects and publications:

Behavior Change, Sustainability, and Climate Communication

We examine how AI can be used (and misused) to shift attitudes and behaviors, and where the gap lies between AI-predicted persuasion and what actually changes in real people.

Selected projects and publications:

Beliefs, Trust, and the Placebo Effect of AI

People's prior beliefs about AI shape what they perceive AI to say and do. Our work shows that priming users' expectations can significantly change perceived trustworthiness, empathy, and effectiveness of the same underlying model, and that users often place misplaced trust in AI even in high-stakes settings such as medicine.

Selected projects and publications:

​