By Roy F Rada, MD, PhD
Placed on LinkedIn: June 13, 2024
Keywords Government, Peace, War, Artificial Intelligence, Online Games, MMOG
What’s Going On. Marvin Gaye July 1980 Rotterdam. Rob Verhorst/Getty Images
Lyrics from "What's Going On":
Father, father
We don't need to escalate
You see, war is not the answer
For only love can conquer hate
You know we've got to find a way
To bring some lovin' here today …
Executive Summary
Imagine a multiplayer online game (MOG) that incorporates artificial intelligence and that has an agenda of world peace. The players enter a MOG and choose an avatar replete with code that runs the avatar when the player is not in the game. The player assumes roles in a family, a company, a government, and other organizations.
Players earn money by growing organizations. Organizations are based in states that have taxing and military powers. However, the game rules forbid war as enforced by a United-Nations-type entity called the The Inner Sanctum. The states have representatives to the Sanctum and democratically legislate its actions. As every entity in this game must consume energy to survive, seeking new sources of energy is a constant challenge, and robots are sent to uninhabited regions to mine resources. The Sanctum has the special privilege to interact with Reality.
This article is Part 6 in a series of articles about AI and building a better world that has addressed autonomous systems, artificial organizations, robots in space, AI in enterprise systems, and game ethics. This article next sketches requirements of the MOG. Finally, a plan to fund R&D is presented.
My previous article “Souls, Suffering, and Online Games”
described how an online game has a creator who defines certain rules which can only be changed by the creator and
speculated about opportunities and responsibilities for an online game to support world peace.
My friend Hafedh Mili’s comment on that article was:
With the advent of remote and virtual reality warfare, I had imagined the opposite of what Roy is suggesting: I imagined a dark future where gamers are enrolled by governments, through slick MMOGs, to fight real wars, without their knowledge! Wars can be traced to a combination of a) disregard for human life, b) inaccurate reading of the adversary, and c) erroneous cost/benefit analyses. To what extent can a MMOG help with these factors?
This article proposes a game that answers Hafedh’s question in the affirmative. The game is a simulation of society built by the players with the agenda of contributing to world peace. My hope is that the game has some merit, but my concerns include that the game is too abstract, too work-like, and outdated.
Roy Rada (left) and Hafedh Mili (right: thirty years ago
I am working on this problem of a MOG, AI, and Peace because I want to contribute to a better world. My PhD thesis half a century ago was an attempt to simulate intelligence in a cellular automaton by beginning with a primordial broth. I got nowhere. Half a century later I was astonished by ChatGPT. I knew intimately neural networks and document libraries but did not anticipate bypassing millions of years of evolution by training a neural network to predict the next token.
ChatGPT lacks agency -- has no sense of purpose. However, a tool like ChatGPT can be combined with other tools to create agency. In the 2023 paper “Generative Agents: Interactive Simulacra of Human Behavior” https://arxiv.org/abs/2304.03442, the authors describe an online game in which artificial people interact with one another and with human players in a way that manifests believable
individual behavior and
emergent social behavior.
Their architecture for each person has three modules:
Long-term memory which records, in natural language, the agent’s experiences and retrieves and combines relevant, recent records to inform the agent’s moment-to-moment behavior,
A reflection module that synthesizes memories into higher-level inferences over time to guide behavior, and
A planning module that translates the reflections into high-level action plans and then recursively into detailed behaviors.
In the past two years, many published papers present various multi-agent architectures using generative AI to create life-like online games.
The 1993 winner of the Nobel Prize in Economics Douglass North said:
Institutions … structure political, economic, and social interaction. They consist of both informal constraints (… codes of conduct), and formal rules (constitutions, …). …. They evolve incrementally, connecting the past with the present and the future; history in consequence is largely a story of institutional evolution …
Could millions of people playing an online game create an artificially peaceful society and from that contribute to peace on earth?
Douglas North in the bottom row far right at the Nobel Prize Ceremony 1993
A person is a living system (LS). A LS has a boundary through which energy and information enter and exit. This energy and information are transported to and from a processing center that bi-directionally connects with storage and is controlled by a decision process. A growth mandate guides the decision process.
Flowchart Characterizing a Living System
People give rise to non-organic LSs from groups (e.g., family) to organizations (e.g., Microsoft) to states (e.g., USA) to a supra-state (e.g., United Nations). Each LS consumes resources proportional to its complexity and has a:
charter which specifies its vision and objectives and
network of roles in which each role has functions, and each function inputs certain messages, operates on the message, and outputs other messages.
At startup, our game is pre-populated with LSs and reusable libraries of components. The game designer can decide the startup specifics. For instance, if the game begins with chiefdoms, then the chief:
declares himself god,
taxes members of his tribe and requires some to serve in his military, and
coordinates services for his tribe.
While the game may be set in any historical period, all LSs will have access to the knowledge reflected in state-of-the-art LLMs.
Players will need to make decisions about which roles in which LSs to join. For instance, a student decides whether to get a job or continue education. The game will offer tools to help players weigh the costs and benefits of decisions. For instance, data models created from a tool, such as Microsoft Power Business Intelligence, will support
charts, data visualizations, and dashboards and
what-if scenarios.
Since the LSs details of operation are explicit, analytics and visualizations readily apply.
LSs strive to create order from disorder. To this purpose they must increasingly consume energy. Our planet currently supports 195 states. Inevitability imbalances occur, and a state sees an opportunity to use war to gain resources. The number of states is too small and the resources too limited, for the dynamics of evolution to produce a higher level of organization than the state. One approach to this dilemma is to find new resources and create a larger evolutionary platform from which to evolve a newer, better kind of state.
With AI and robotics, the LSs send robots to otherwise uninhabitable regions where the robots gather resources otherwise ignored – this is the Last Frontier. Examples of the Last Frontier are the ocean floor, the polar ice caps, and outer space. These regions provide untapped energy and material reserves and other strategic opportunities. The US Space Force has said (https://www.defense.gov/News/News-Stories/Article/Article/3723145/space-force-general-outlines-us-approach-to-maintaining-space-superiority/): "Space’s critical role in enabling terrestrial operations requires that the US always maintain space superiority."
an Inner Sanctum level (the term Inner Sanctum is a colloquial form of Sanctum Sanctorum which is Latin for Holly of Hollies).
The Sanctum monitors the behavior of states. It takes by fiat a small portion of the taxes collected by the states and has the authority to issue orders to each state’s military. The Sanctum has a legislative division composed of representatives from the states.
Vladimir Zelensky addresses the UN Security Council February 2024
States in the real world continually oscillate between peace and war. In our game, war is not allowed -- the Sanctum prevents states from going that far.
LSs at the regular level are confined to the game world, but the Sanctum may transact with the real world. These external transactions include:
consulting to spread peace in the real world and
acquiring resources with robots deployed to uninhabited places.
Consulting for world peace would be based on experiences gained in the online game. Robots are launched when:
they will earn more than they consume and
Sanctum negotiations with real states assure cooperation.
With extra resources, the Sanctum bargains for peace.
Robot on the moon collecting resources and strategically positioning itself
Our game is a sandbox that simulates society (think SimCity or Second Life). Players enter the game and assume various roles. The players' two quests are to:
earn money and
contribute to world peace.
They pursue these quests by building LSs and deploying robots.
Screenshot from SimCity produced by Electronic Arts 2024
The game encourages reuse. Libraries of reusable components are provided. LSs may also acquire other LSs or merge with them.
Players on entering the game will assume an avatar that is a 'game person'. The game person is continually functioning in the game based on explicit programs that are processing input and generating output. However, when the human is in the game, the human can over-ride or re-write his game player program. The game person may assume roles in LSs, such as the role of a family member, an employee of a company, or a student at a school.
The ideas presented in this article are not a business plan but brain-storming for a R&D initiative. I would like to create an entity, call it Funder, which solicits proposals from universities to explore the ideas.
The Funder will seek to partner with established institutions to host the Fund. The Funder has a small startup grant and the earnings from that grant support the pump-priming efforts of the Funder. After partner(s) are established, the Funder might
make small grants for pilot projects and
seek funding from external sources.
The Funder will operate as a virtual organization. Its website will publish
its charter and
the requirements for proposals to the Funder.
The Funder wants to help university professors and students create MOGs that support the peaceful growth of organizations.
This article is part of a series, and that series explains how AI can bring organizations to life and how this can be done mindfully.
