AI that makes people harder to exploit
We build for a world in which intelligence increases people's power to understand, decide, refuse, cooperate and leave. The useful work, knowledge and gains from automation should stay with the people and communities whose lives create them.
That is the 561 Group's position on aligned AI. We want technology that weakens exploitation, breaks compulsory dependence and expands the means of living well together. When that undermines a profitable business built on captivity or artificial scarcity, we regard the loss of that business model as a reason to keep going.
We intend to make this future through research, software, institutions, public services and places where people can practice it. Our business earns its place by helping people become more capable.
Alignment is a political choice
A system's purpose shapes whose life becomes easier. An advertising system can make attention more profitable to capture. A workplace system can turn every minute into a demand for more output. A platform can use its intelligence to make leaving harder. The people governed by those systems need the power to change their purposes.
We take the other direction: knowledge that people can use, tools they can understand, authority they can withdraw, work they can retain, and institutions they can reshape together. Aligned AI should weaken capitalism's power over those capacities. It should make it easier to meet needs, share knowledge, coordinate useful work and refuse extraction.
This is a question of ownership, maintenance and collective life as much as model behavior. Who sets the task? Who can challenge it? Who performs the checking and repair? Who keeps what the work creates? Our research and our commercial choices begin with those people and relationships.
The wider discussion is catching up with these questions. O'Neill and colleagues' economic-alignment research examines growth incentives, commons and tools that enhance autonomy. We bring the Group's relational philosophy and agentic computing research to that conversation.
From Relationality to a living Web
The Group is Lakota-founded. Its founder's Relationality develops into Obsmetrologia, mathematical and computational work, and the design of agents and institutions. Relations are the starting point: what something can do, how it changes, and what its activity means arise through its relationships.
That gives our computing a demanding purpose. A system must be able to encounter something that changes its understanding of the situation. Its memory, categories, commitments and actions must develop together. People need to participate in shaping those relations, including the power to refuse them.
Our Jiangshi Web argument describes what happens when recorded form and automated activity continue while the capacity for reflective change is cut off. In the essay's Jing, Qi and Shen vocabulary, Jing names stored structure, Qi its circulation and activity, and Shen the capacity to recognize and revise the frame through which a situation is understood. More activity can keep the body moving while its joints remain locked.
An agentic Web can deepen that condition or transform it. We pursue systems whose categories and commitments can change through situated interaction, carrying those changes into memory, decisions and action. Read the Jiangshi Web and SHRIMP research at Lenticule Science.
Black October brings the problem into view
The Black October incidents expose different breaks between knowledge and action. The PHP Foundation describes a repair delayed by difficult compatibility decisions and insufficient specialist resources. LiteLLM's investigation traces its compromised release through an exposed key that remained active and excessive repository authority.
Faster discovery increases the importance of the work around it: deciding, maintaining, revoking, repairing and helping people receive the change. These are central subjects of the Group's research, engineering and operating practice. We want the benefits of greater intelligence to reach the people living with the systems it changes.
What we build for
| Principle | Practical expression |
|---|---|
| People can change what a system relies on. | A correction travels through the explanations, decisions and actions that depend on it. Disagreement remains available for thought and further work. |
| Delegated authority remains answerable. | People can withdraw permission, stop affected actions and understand what continues under the authority still given. |
| Useful work stays useful after exit. | Results, necessary context and agreed rights travel with the person or community that needs them. |
| Automation increases people's capacity. | We account for the whole task: doing, checking, repairing, maintaining and living with the result. |
| Communities govern their knowledge and participation. | Collective benefit, authority to control, responsibility and ethics shape how systems are built and used. |
The CARE Principles for Indigenous Data Governance articulate that last relationship. Collective authority belongs in the operation of technology: in who can decide, what can be shared, how withdrawal works and who benefits.
These principles give collaborators concrete things to work on with us. Bring an agent whose authority is difficult to withdraw, a result that is trapped in a platform, or a correction that never reaches the people relying on it. We can turn that problem into a clear research question or a useful piece of work.
A business worth choosing
We earn through research, implementation, maintenance, teaching, hosting and services that people find useful. Continued payment should follow continued usefulness. A customer who gains enough capability to stop needing an engagement can be a success.
A website engagement should leave its owner with a working task and a usable handover. Editorial work should leave an explanation the client can understand, maintain and reuse. An automation engagement should establish who directs it, who can stop it and who retains the output. Each engagement makes its deliverables, rights, costs and support clear.
Revenue supports the people, maintenance and shared infrastructure that make this work possible. We want that economy to grow through the capabilities it gives people and the relationships they freely choose to sustain.
Cultivate the conditions for something better
Cultivation changes what can happen next. A useful explanation makes later inquiry easier. A well-maintained tool supports work its maker could not anticipate. A community becomes more capable through the relationships, knowledge and shared practices its participants develop.
This is why we cultivate attention on the Web. We want an encounter with the Group to leave someone with a stronger capacity to think and act: a source worth remembering, a question worth pursuing, a tool worth using, or people worth working with.
Turntide Islands is a place to take part in shared inquiry, learning and adaptation. Our worlds, games, stories and public learning are part of the life we are building. Play and collective imagination help people discover possibilities together.
Read the research, follow one correction, or choose a task in From research to a useful result. Use the free business worksheet or Web task worksheet to begin now.
Bring us a place where people could have more power to understand and act. Discuss it with Causeway Services, explore the work, or join the inquiry. That is how we intend to build a different future.