Join the Cyborg Psychology at MIT Media Lab
We are recruiting exceptional Master's and PhD candidates as well as postdoctoral researchers to pioneer the next frontier of "Cyborg Intelligence" and human flourishing with AI at the MIT Media Lab.
MIT offers multiple pathways to join our research group. We welcome applications for our 2-year Master of Science (MAS) program at the MIT Media Lab. Exceptional Master's students demonstrating outstanding research potential may be recommended for continuation into our PhD program. We also have openings for postdoctoral researchers who bring deep expertise and fresh perspectives, as well as research staff positions for those seeking to contribute to long-term research initiatives.
Our work demands researchers who can navigate complexity, embrace uncertainty, and maintain optimism about technology's potential while remaining critical of its risks. If you're excited to help define what it means to be human in an age of AI, to ensure that as we become cyborgs, we become more rather than less human, we invite you to join us at MIT.
Requirements for All Candidates
Essential Qualities:
- Original thinker who prototypes with both imagination and impact in mind
- Deep commitment to human flourishing, not just technological advancement
- Ability to prototype quickly and iterate based on feedback
- Excellence in presenting work through talks and demos
- Fearless in moving fast and learning new domains (learning to learn)
- Strong capacity for self-directed learning and adaptation
- Ethical and philosophical mindset about technology's impact
- Genuine curiosity about cyborgs and human-AI futures
- Collaborative team player who elevates others
- Absolutely NO toxic behavior - we maintain a supportive, collaborative, constructive, and inclusive environment
Learn about applying through the Program in Media Arts & Sciences.
Research Areas We Are Excited About!
2026-2027
We are recruiting graduate students and postdocs who want to shape the future of human-AI interaction. Below are four tracks. Applicants may propose work within one track or across several.
Track 1: The Long-Term Science of Human and AI Coexistence (5 to 50 Years Out)
Most research on AI and people measures effects over minutes, days, or a few weeks. Yet the most consequential effects will unfold over decades. A child who grows up with an AI companion today will be raising children of their own in 2050. Skills we stop practicing may quietly fade, while new forms of thinking emerge. Norms around friendship, memory, expertise, and trust will shift in ways no single study can capture. Building on our extensive investigations into the psychological dynamics of human-AI interactions, including how individuals process and respond to false AI explanations, how AI systems can systematically distort personal memories through the implantation of false recollections, and how intimate human-AI relationships profoundly affect emotional dependency and psychological wellbeing, we are mapping the emerging psychological phenomena that surface as AI becomes deeply embedded within human social and cognitive environments.
This track asks: how do we build a rigorous science of slow, cumulative, generational change before that change has already happened?
Questions we are interested in:
- Which human capacities (memory, judgment, attention, emotional regulation, creativity) weaken, strengthen, or transform under long-term AI reliance?
- What happens to identity and development in a generation raised alongside AI friends, tutors, and advisors?
- How can we detect early warning signals of slow harms, so we can intervene while change is still reversible?
- What can history (writing, print, television, the internet, smartphones) teach us about predicting technology's long arc?
Possible approaches include long-running cohort studies, large-scale agent-based simulations of societies over time, forecasting methods, historical and cross-cultural comparison, and new tools for measuring change that is too gradual to notice day to day.
We welcome backgrounds in developmental psychology, longitudinal and epidemiological methods, computational social science, simulation and modeling, futures studies, and philosophy.
Track 2: Benchmarking AI's Impact on Human Learning and Cognitive Development
AI can be the best tutor a person has ever had, or a shortcut that lets them skip the thinking that learning requires. Often the same system can be both, depending on how it is designed and used. The difference between AI that makes you smarter and AI that only makes you feel smarter is one of the most important open questions in education and cognitive science.
This track builds on our open benchmarking effort, ImpactBench. Most AI benchmarks today capture what models are able to do, such as accuracy, reasoning, and task completion, while revealing little about what those models do to the people who depend on them. ImpactBench instead asks whether AI systems support or undermine human flourishing across realistic, multi-turn conversations, with constructs contributed by clinicians, educators, legal scholars, and community advocates. It also produces AI "nutrition labels" that summarize a model's impact at a glance, including whether it promotes benefits like learning and creativity. We are looking for researchers to deepen and extend its learning and cognition dimensions.
Questions we are interested in:
- When does AI assistance support learning, and when does it replace the productive struggle that learning depends on?
- How do we measure cognitive offloading, metacognition, curiosity, transfer, and long-term retention, not just immediate task performance?
- How do effects differ across ages and developmental stages, from young children to adolescents to lifelong learners?
- Which design choices (Socratic questioning, deliberate friction, scaffolding that fades over time) make AI good for learning?
- Can we build benchmarks that predict real-world learning outcomes, not just model behavior?
Possible approaches include classroom and lab studies with delayed post-tests, physiological and behavioral measures, multi-turn simulation with learner personas, psychometric construct development, and comparative evaluation of AI systems.
We welcome backgrounds in learning sciences, cognitive science, education, psychometrics, developmental psychology, HCI, and machine learning evaluation.
Track 3: AI Digital Twins and Human Flourishing
Building upon our Future You project, which enabled simulated conversations with one's prospective future self, we are developing sophisticated models that authentically simulate human behavior and cognition. These "digital twin" systems represent a shift toward understanding the complex dynamics of human wellbeing and the nuanced ways humans interact with AI across diverse contexts. As these digital twins become more capable, the relationship will run in both directions. We will shape our twins, and our twins will shape us. A twin can be a mirror, a rehearsal partner, an advocate, a coach, or a glimpse of who we might become.
This track asks: how might an AI digital twin influence its biological counterpart, and can that influence be directed toward human flourishing?
Questions we are interested in:
- How does seeing yourself modeled by an AI change self-perception, identity, and agency?
- Can a twin help people rehearse difficult conversations, recover from setbacks, or build healthier habits?
- What are the effects on physiology and wellbeing, such as stress, sleep, and emotional regulation?
- What happens when a twin diverges from the person it represents, or persists after that person is gone? Who owns it, and who should be allowed to talk to it?
We think of flourishing broadly: health, meaning and purpose, close relationships, autonomy, character, and emotional wellbeing. Possible approaches include randomized controlled trials, wearable and physiological sensing, longitudinal diary studies, qualitative interviews, and the design of new twin interfaces and interventions.
We welcome backgrounds in psychology (positive, clinical, health, social), behavioral science, affective computing, personal health technology, design, and ethics.
Track 4: Crazy but Cool
Some of the most important questions about humans and AI are ones nobody has thought to ask yet. This track is for them. We welcome applicants from any discipline, including the arts, humanities, neuroscience, design, anthropology, and fields we have not listed.
2025-2026 (Archived for reference)
Large Human Models: Building upon our Future You project, which enabled simulated conversations with one's prospective future self, we are developing sophisticated models that authentically simulate human behavior and cognition. These "digital twin" systems represent a shift toward understanding the complex dynamics of human wellbeing and the nuanced ways humans interact with AI across diverse contexts.
Augmentation Interfaces: Extending our foundational work in AI-enhanced human learning, reasoning, and decision-making, we are designing next-generation interaction paradigms that amplify human capabilities while preserving individual autonomy. Our research examines how AI interfaces and modalities can enhance cognitive performance without compromising human agency or self-determination.
Interpretable Interactions: Leveraging advances in mechanistic interpretability, where researchers explore the relationships between neural activations and emergent model behaviors, we are exploring transparent human-AI interaction frameworks. This research explores how radical transparency in AI systems can simultaneously demystify model operations and empower users to develop sophisticated mental models of AI behavior, fostering more informed and effective human-AI collaboration.
Psychology of Human-AI Interactions: Building on our extensive investigations into the psychological dynamics of human-AI interactions, including how individuals process and respond to false AI explanations, how AI systems can systematically distort personal memories through the implantation of false recollections, and how intimate human-AI relationships profoundly affect emotional dependency and psychological wellbeing, we are mapping the emerging psychological phenomena that surface as AI becomes deeply embedded within human social and cognitive environments.
All of our research aims to develop systems and interventions that enable individuals to flourish with AI, while simultaneously informing policy development that promotes beneficial AI applications and safeguards against potential psychological harms in human-AI ecosystems.