Plural Technology Collaboratory
Building the technological foundations for a plural society, where diverse communities can cooperate, self-govern, and thrive through advanced AI.
Why it matters
The Plural Technology Collaboratory is a project of the Microsoft Research Catalyst Lab. We ask an ambitious question: as foundation models reshape the world, how can free, diverse, and democratic societies flourish? We believe this deserves research at the same scale as building the models themselves.
The world is investing enormously in making AI more capable, and only a small fraction of that in the civic and institutional architecture society needs to absorb it. We work on that missing half: the sociotechnical systems that help communities keep pace with digital innovation.
Our approach was pioneered in Taiwan from 2014 to 2024 under its first Digital Minister, Audrey Tang, and articulated in Plurality: The Future of Collaborative Technology and Democracy (opens in new tab), the first open-source, democratically governed book, endorsed by His Holiness the Dalai Lama (opens in new tab) and a bestseller in Japan (opens in new tab).
How we work
We move from theory to technology to real-world deployment, combining political economy, cryptography, and human-computer interaction to build systems and study their consequences with real communities.
Our areas of impact
Sociotechnical Grand Challenges
The world is spending vast sums to build ever more capable AI, and far too little on the social and institutional changes needed to live well alongside it. We call these the sociotechnical grand challenges of the AI age: hard, multisector problems like identity and provenance, data governance, democratic innovation, and the reinvention of education. Solving them will take coalitions across technology, government, civil society, and faith communities, not consensus alone. Our framing essay, “What Humanity Needs To Flourish In The Next Decade”, lays out this agenda together with a coalition of dozens of collaborators across sectors, countries and perspectives.
Communal and organizational alignment
Most AI is trained to speak in a single, flattened voice or to be sycophantic to users. We are developing a different approach, communal and organizational alignment, in which AI represents the full range of relevant human perspectives and helps surface where communities and organizations agree and where they differ. This is the logic behind systems like Community Notes and Taiwan’s vTaiwan, and it points toward media and AI that strengthen the social and organizational fabric rather than strip-mine it. Our paper “Community by Design” and our Noema essay “Building a Prosocial Media Ecosystem” set out how to build it in the social media context and we are extending this work to AI interactions.
Data dignity
To sustain itself and grow, the AI economy needs to reward and empower those who create the data on which it runs. This idea was called “Data Dignity” by Satya Nadella in 2018 and builds on our work in the Harvard Business Review with Office of the Chief Technology Officer Prime Unifying Scientist Jaron Lanier in 2018, “A Blueprint for a Better Digital Society”. Together with Raul Castro Fernandez of the University of Chicago, we develop a detailed, technically-grounded practical design for a market that, at essentially no additional computational cost, pays fair rates for AI training data, described in another HBR article “How AI Companies Can Pay Fair Rates for the Content They Need”.
Civic and democratic technology
To survive the arrival of AI, democracy needs new digital infrastructure that helps people deliberate, find common ground, and shape the decisions that affect them. We study and support participatory civic technologies that put this into practice: Taiwan’s vTaiwan and Join platforms, Pol.is-based deliberation, Alignment Assemblies that bring citizens into decisions about AI, California’s EngagedCA platform and the German government’s Deutschland Wie Geht’s and, most recently, Team Mirai in Japan, a technology-driven political movement that used AI to listen to voters at scale and won seats in the national legislature. Across these cases, technology is used to widen participation and bridge divides rather than to concentrate power.
Technology for Religious Empowerment (T4RE)
Faith communities are among the oldest and most resilient institutions for helping people live by their values, and they are often overlooked in technology design. Technology for Religious Empowerment partners with religious communities across many traditions to build and evaluate tools that strengthen, rather than disrupt, communal and spiritual life. Our work includes a landscape study of stakeholders across many faiths, the Good Faith Design initiative, AI red-teaming with faith leaders, and engagement with the Faith, Family and Technology Network.
Our external collaborators
We are working closely with the following collaborators outside Microsoft, as well as a variety of internal groups:
