Companions & Consequences
The texture of intelligence

Review of the MIT Media Lab Symposium on
Advancing Human Enhancement with AI

The inaugural MIT AHA Symposium – titled “Can we design AI to support human flourishing?” – took place in April 2025 at the MIT Media Lab. This one-day event launched the Advancing Humans with AI (AHA) initiative, focusing not on smarter machines, but on better human outcomes. The symposium grappled with AI’s double-edged impact: its potential to enrich lives versus risks like overreliance, lost skills, and social disconnection. With AI now poised to mediate everything from private musings to public discourse, the forum urged a proactive, human-centered framing of our relationship with intelligent machines.

Inner Life and Artificial Intimacy

Today, people spend more time alone with technology, than ever before. AI, once a tool for cognitive tasks, now approaches the emotional core of our lives—relational-AI, designed to simulate conversation, care, even companionship.

MIT professor Sherry Turkle warned we may be entering a “robotic moment,” when machines are not only accepted as social actors, but invited into intimate emotional spaces—especially by children who have yet to experience the full depth of human relationships. Anthropomorphic systems may offer the form of intimacy while withholding its substance.

Such systems create what we might call artificial intimacy—an experience that mimics connection, but lacks the moral presence or mutual vulnerability that defines genuine relational bonds. Is it ethical to design machines that pretend to care? Do we risk confusing responsiveness with empathy?

As historically noted by McLuhan, the medium is the message. When AI becomes the medium through which we relate, it reshapes the relational field itself. The consensus: AI should support—not supplant—our social fabric. It may assist, mediate, even enhance—but it should not impersonate.

Enhancement vs. Substitution

A different thread explored whether AI should be framed as a collaborator or a substitute. Will it elevate or displace our capacities and agency?

To navigate this, one group proposed a new conceptual tool: the EPOCH Index—a human-centered metric evaluating the extent to which a task draws upon empathy, moral reasoning, creativity, and other distinctly human faculties. Paired with automation and augmentation metrics, the index helps locate where AI might amplify human strengths without hollowing out their practice.

The late Ivan Illich's call for 'convivial tools' cast a light on the event—technologies that, rather than breeding user dependence and impotence, in fact, serve user autonomy, and foster social creativity. “If you build a machine to help you think, it may also begin to think for you.” The ethical design of AI, in this view, demands limits: augmentation without alienation.

This distinction echoed throughout the day. Jaron Lanier (VR Pioneer and 'Prime Unifying Scientist' at Microsoft) emphasized the importance of process over output: when AI supplies answers too easily, we may lose the generative tension of trial and error, the formative effort behind insight. As with the microscope, AI should be an instrument that scales thought, not one that overrides it. Jaron's views on AI and "data dignity" seriously question current AI hyperbole. With reference to his New York Times piece, "There is no AI", in his view AI is most powerful as an innovative form of social collaboration. That it is not at all an independent entity, but in fact, made by and therefore to be viewed as wholly inclusive of, people.

Learning and Knowledge: Orchestrator or Oracle?

Nowhere is the human-AI threshold more pronounced than in education. Can AI support the act of learning, or does it risk trivializing it?

Educators at the symposium described the seduction of the shortcut: instant answers delivered by intelligent tutors. But the act of learning is not merely informational—it is relational, temporal, and often difficult. It is through struggle and discovery that understanding takes root.

AI, they argued, should be designed as an orchestrator of knowledge—guiding learners toward resources, nudging them with prompts—rather than as an omniscient oracle. This repositions AI from authority to companion, emphasizing agency over automation. The goal is not efficiency alone, but engaged, meaningful learning that remains anchored in human cognitive development.

Metrics and Vision: A New Frame for Progress

The symposium’s most ambitious contribution may be conceptual rather than technical: a redefinition of progress.

Instead of measuring AI’s success by scale, speed, or predictive power, we might ask: Does it expand human possibility? Does it deepen our collective agency? This reframing invites new metrics—not just performance benchmarks, but indicators of empathy, connection, creativity, and communal wellbeing.

The AHA initiative is already piloting several moonshot projects to make this vision actionable:

Atlas of Human-AI Interaction: mapping patterns of engagement between people and AI systems.
Benchmarks for Human Flourishing: developing new standards to assess how AI impacts agency, creativity, and connection.
Global Observatory for Human Impact: monitoring the ethical and cultural implications of AI across diverse regions.

Together, these initiatives suggest a future in which technology’s value is indexed to human vitality—not simply what machines can do, but how they support human flourishing.

The Texture of Intelligence

If AI once promised us efficiency, it now invites us into something more delicate: a negotiation between presence and performance, between simulation and solidarity. The texture of intelligence—its warmth, rhythm, responsiveness—cannot be reduced to pattern recognition or optimization. It must be cultivated.

Designing AI for human flourishing is not just a technical problem. It is an aesthetic, social, and philosophical endeavor. And it begins with a deeper question: What kind of world do we want to live in—and what kind of beings do we want to become?

Media Details

Details on MIT Media Lab's - Advancing Humans with AI - Multi-faculty research program can be found here:

https://www.media.mit.edu/posts/introducing-aha/

Period: 2023-2025 | Language: English
Core Concepts: Relational-AI, Human Flourishing, Artificial Intimacy
Audio-Visual Media: c/o 10 Sensor Concept
AI-Usage: Generative AI, source & output validation, model-switching
Conflict of Interest: None
References: MIT Media Lab, AI & Human Advancement Symposium, 2025 | "There is no AI", Larnier, The New Yorker, 2023 | STEM Working Groups 2024-25 | MIT AHA Community, 2025

Back to Futures