If we want new technologies to increase humanity’s collective intelligence, we must do two things: (i) facilitate many natural experiments, and (ii) restore the epistemic independence of our dialectic communities.

intellectual background

New digital technologies are rapidly changing the way we collaborate to make sense of the world. These include: social media technologies, which change how we share information with each other; AI and digital search technologies, which change how we draw insights from text and data; and collaborative editing technologies, which change how we jointly complete tasks.

These technologies can enhance humanity’s collective intelligence, but the path might not be smooth. Gutenberg’s metal moveable type printing press provides an illustrative example. The press eventually set the stage for the enlightenment and the scientific revolution. However, early mass-printed works, such as Malleus Maleficarum, rapidly spread lies about witches. Because books were previously so rare, humanity had not yet adopted norms or refined institutions to mitigate the impact of this misinformation. The resulting “witch crazes” led to between 20,000 and 60,000 brutal deaths. Over time, humanity refined institutions and adopted new norms, allowing us to mitigate the downsides of printed lies while still enjoying the manifold benefits of the press’s information revolution.

Humanity is still refining the institutions and developing the norms that will allow us to more fully realize the benefits of our new digital technologies. This search process is slow, due in part to the relationships between the technologies’ affordances, and our norms and rules. The impact of any particular affordance on our collective sense-making depends on more that one affordance. It also depends on rules, norms, and other affordance decisions. For example, the impact of an affordance that allows people to edit each other’s online comments would greatly depend on whether social norms encouraged or prohibited individuals from misrepresenting their adversaries’ positions.

While the above toy example is extreme, there are many such “epistatic linkages” or “interdependent relationships” between norms, rules, and digital affordances. These interaction effects mean that there are many local optima on the solution landscape that maps each set of norms, rules, and digital affordances to the expected efficacy of the resulting collective intelligence. We call solution landscapes with many local optima "rugged” or “complex.” If we simply make incremental improvements to one or two initial experiments, we are almost guaranteed to “get stuck” on a sub-optimal (local) equilibrium when searching a rugged landscape. In order to search a rugged landscape, we need to mix rigorous theory with many experiments.

Fig 1. A rugged landscape.

the ARCH solution

Enabling diverse natural experiments, and ethically collecting comparable data from those experiments, is the key to identifying sets of affordances, norms, and rules that promote high-quality human collective intelligence. To promote such experiments, Project ARCH is building a highly-customizable, open-source, collaboration platform for dialectic communities. At present the ARCH platform is supporting scholarly collaborations, but we will soon move to support entire colleges and universities. Eventually we will also offer our platform to churches, unions, clubs, political organizations, and other civil institutions.

The platform offers text chat, video chat, virtual lectures, videos, discussion threads, collaborative documents, and AI-tools. Administrators of the individual networks can toggle features on and off, and moderate the affordances to best support their community’s goals. Network administrators also have control over the data generated by activity on their network, but the data is structured to enable cross-network comparisons when the administrators agree to particulate in such studies. We are presently working with IRB at top universities to ensure that data collected on the platform can used in work intended for peer-reviewed journal publications.

We are currently focused on helping scholarly communities deploy digital tools that enhance their collective intelligence and strengthen their cohesion. These communities are most often inter-university groups of scholars and practitioners working on specific problems. The ARCH platform is often launched in conjunction with a workshop or working group, where the organizers are trying to build a more sustainable community of inquiry. We also currently work with summer schools targeting graduate students, postdocs, and practitioners. Organizers will often ask us to stage their ARCH Network - so we turn on different tools as their community grows and evolves. We actively share lessons across the communities we support, to enable broad and intentional search of how different mixes of digital tools impact the community’s collective intelligence.

Thanks to generous support from the Sloan Foundation and the Omidyar Network, ARCH is free for scholarly communities. We currently host all ARCH Networks on our own servers, but we expect to offer our partners the option of self-hosting their ARCH Network with open-source ARCH software in the next year. If you are interested in using ARCH to support your own intellectual community, please let us know. In fall 2026 we plan to offer a school-wide ARCH solution for colleges and universities, if you would to be informed when this option becomes available, please let us know.

Workshop

ARCH can support in-person meetings or ongoing communities of intellectual inquiry.

Collaborate

Host virtual conferences, edit collective documents, and more.

AI Tools

Experiment with different deployments of AI to find solutions that enhance your collective's intelligence.

Archive

Store papers, videos, and links in the built-in repository.

Fig 2. The logistic map traces an ARCH.

benefits of project ARCH

Intellectually, ARCH helps humanity explore the relationship between digital affordances, norms, rules and the efficacy of the resulting collective intelligence. However, there are at least two other (more applied) benefits:

ARCH helps the administrators regain a degree of epistemic independence over their communities. Increasingly, community leaders in academia, churches, unions, clubs, political organizations, and other civil societies, rely on commercial platforms to connect their community. This means the digital affordances that enable communication and collaboration in these civil institutions were largely selected to enhance engagement and the collection of monetizable personal information. ARCH allows administrators to select the digital affordances that best support their communities’ goals. This will improve these communities’ collective intelligence.

Finally, ARCH is being designed to use open protocols which will allow users to blend content from the different ARCH networks to which they belong. We believe rapid advances in AI technology will soon make this task both easier and more efficient. This change will increase the community structure of the ad hoc network through which democratic societies collectively make sense of the world. There is reason to believe that adding this structure to the society-wide, ad hoc dialectic network will make us more resilient to foreign misinformation campaigns, and more effective in our domestic deliberations.

As an open source project, many of our collaborators balance their work with other obligations. Our bottom-up team is constantly growing. Please let us know if you would like to be involved.

advisory board