I would like to propose a new way of thinking about communities and how they are evolving, with the advent of AI and Large Language Models (LLM’s).
Previously I have written about how group minds are emerging, facilitated by LLMs. Here I will delve a little further into how we can look at types of communities on a spectrum, where the final stage is communities of intelligence, which we are entering now.
Below is how I view the evolution of communities by stage, with some comments.
LEVELS OF COMMUNITY
1. Communities of Situation – defined merely by being adjacent in space and/or to time. For example, being at the same geographical place, or participating in the same event.
2. Communities of History – defined by a common history – such as by having shared experiences, shared historical relationships to each other, events, familial links, or relationships to a third party, etc.
3. Communities of Interest – defined by set of shared interests, relationships, or needs in the present and future.
4. Communities of Practice – defined by set of shared skills, expertise and experience.
5. Communities of Purpose – defined by set of shared goals, processes, or projects
6. Communities of Contract – defined by participating in an shared agreement, such as a social contract, legal contract, etc.
Citizens of locales are participating in multiple communities of contract for their city, state, nation, etc. Similarly participants in a corporation, or a business deal, or any legal contract, are in a community of contract for that set of relationships, history, shared interests, etc.
7. Communities of Attention – defined by set of shared foci.
For example, Twitter is a community of attention. The community as a whole, as well as sub-groups within it, acts as a hierarchy of attention economies around trending topics, hashtags, and people. The result of this activity is a set of filters for focusing the attention of the audiences.
In a community of attention, the community is both the sensing mechanism (e.g. what is being measured to gather information about what is being focused on), and the consumer of the reports that are generated from that data (e.g. such as lists or maps of trending topics, trending messages, trending people, etc.).
8. Communities of Intelligence – defined by participation in a set of collective intelligences, which may include AIs.
Here what binds participants is that they are part of the same collective intelligence – which means they both contribute to and consume the intelligence.
Communities of intelligence are sets of people and AIs that share common collective intelligence – a pool of specific people and AIs trained by and/or for these people, and whom intelligently assist one another.
Consider each community of intelligence as a black box. If you send messages to the black box it responds intelligently, and it does so from its unique perspective – its past and present knowledge and the brains of its people.
Inside the black box we can see that this results from the unique pool of people and AIs that provide its intelligence. Outside the box, we simply see that it represents an intelligence with certain features.
We call the activity of black box of this type, a group mind. A group mind is what emerges from a community of intelligence in action over time. It is the behavior of a community of intelligence.
A group mind is the activity of a community with intersecting situations, histories, interests, practices, purposes, contracts, attention, and intelligence.
I believe that the next stage of evolution for humanity is to participate in group minds. Group minds are what combine humans and AIs together into something more intelligent than either form of intelligence on its own.
We will see group minds emerging as groups, communities and organizations build and grow their own private shared LLM’s to reflect their unique data sets, training, and training biases.
LLMs can be derived by data mining existing community content. For example they could be developed by mining the activity of a community off interest, or a community of attention, such as a social network, over time. They can also be derived by data mining the content of an enterprise.
Long-term I do have some concerns about whether humans have a continuing role in the evolution of intelligence, depending on how powerful AI becomes. It may or may not be the case that future group minds will include humans — they may be completely comprised of AIs as everything the humans know and are capable of is extracted into AIs. Whether this happens, or at least when it happens, is still a topic for study and debate.