Humanity Unleashed

Putting AI in Service of the World

Introduction

1. AI Overlords or Public Servants, the Choice is Ours

2. Social Systems of the Early 21st Century: The Good, the Bad, and the Ugly

3. There's a Better Way: Social Environment Design and the Society of the Future

4. Putting Our Principles Into Action

5. The Path Ahead: Current Work and Roadmap

6. Join Us and Help Take Back Control of Our Collective Future

Appendix

5. The Path Ahead: Current Work and Roadmap

We are currently engaged in building Social Environment Design Version Zero (SED v0). The purpose of this stage of development is to advance the primary technical requirements of the system and align individuals around the construction of SED and the values and beliefs that motivate us toward that end.

Alongside the general tasks needed to facilitate the effort itself, this breaks down into four primary elements which reinforce one another:

1) developing more effective value elicitation mechanisms,

2) leveraging AI for effective and unbiased policymaking,

3) developing incentives for contributions to public goods, and

4) ensuring our impact is positive throughout the process.

Below are the currently completed (checked), underway, and near future tasks within each of these categories:

Alongside the general tasks needed to facilitate the effort itself, this breaks down into four primary elements which reinforce one another:

1) developing more effective value elicitation mechanisms,

2) leveraging AI for effective and unbiased policymaking,

3) developing incentives for contributions to public goods, and

4) ensuring our impact is positive throughout the process. Below are the currently completed (checked), underway, and near future tasks within each of these categories:

Alongside the general tasks needed to facilitate the effort itself, this breaks down into four primary elements which reinforce one another:

1) developing more effective value elicitation mechanisms,

2) leveraging AI for effective and unbiased policymaking,

3) developing incentives for contributions to public goods, and

4) ensuring our impact is positive throughout the process.

Below are the currently completed (checked), underway, and near future tasks within each of these categories:

5.1 General Operational Tasks

Create an operational structure and recruit an initial team for research of SEDv0 under the auspices of the nonprofit company Humanity Unleashed, Inc.

Establish a brand by trademarking “Humanity Unleashed” (granted) and “Humun” (in review by the PTO) (See about trademark of Social Environment Design/SED)

Create a database of related individuals and organizations to reach out to for feedback, collaboration and donation

(Underway) Preparing outreach materials to scale involvement

Conduct an outreach push to garner the involvement and resources needed to scale ongoing research on SED and the system broadly

Scale the team, with a focus on 1) Aligned individuals who believe in the mission and its importance, and 2) Individuals with unique talents and perspectives that can help us develop and communicate the system more effectively

Develop warning signals pointing in either direction of generally positive or negative turns in the broader context

5.2 Developing More Effective Value Elicitation Mechanisms

Review and compile proposed methods for “Democratic Inputs to AI”

Research the most effectual inputs to a system such as SED and the pertinent existing literature within Social Choice Theory

Develop an experimental approach to eliciting foundational values from individuals that allows for their inclusion in the SED process

Draft a related paper

Create a roadmap for enhancing the predictive value of the inputs to outcomes as determined by the participants in the process

5.3 Leveraging AI for Effective and Unbiased Policymaking

Create a framework for tracking contributions toward the creation of the realized SED system

Create the initial guidelines for the system used to reward contributions toward SED (Link to page).

(Underway) Experiment with a simplified SED approach for internal decision making

Draft a related paper

5.4 Developing Incentives for Contributions to Public Goods

Create the theoretical framework for the SED system and AI reward functions

Advance the general framework for “transcendence” that will allow SED’s AI capacities to operate beyond the capacities of the humans directing it

(Underway) Experiment with a simplified SED approach for internal decision making

(Underway) Develop a simplified version of the tracking system and experiment with it internally

Create a ‘contribution portal’ with tracking and attribution allowing individuals to contribute monetarily, through advocacy, or to specific projects we are working on with confidence that the contributions will be tracked and rewarded in the future

Build capacities to ease the management, tracking, reporting and verification processes

5.5 Ensuring Our Impact is Positive Throughout the Process

Volunteer Fridays (Preparing and giving sandwiches to those less fortunate)

Find local opportunities to deploy the SED framework to benefit social causes

Identify and assist in turning around a failing/ineffectual nonprofit

Leverage lessons learned to improve SED (5.3)

Use as a case study to demonstrate the system’s capacities and garner support (5.3)

Donate, ideally directly through a hands on approach

5.6 Putting it All Together

Once we have completed the above tasks, it will be possible to begin to integrate the system and allow for iterative improvement through feedback loops between each element. This will constitute the next phase in the development of SED and will signal that we are ready to launch Social Environment Design Version One (SEDv1). This phase of development will come with new opportunities:

With the aligned organizations and individuals gathered in the previous phase, use the SED system internally to drive its own development
1. Incorporate AI advances for the principal agent and for enhancements of the reward and tracking systems
Motivate individual action through increasing confidence in the system and likelihood of future reward
1. As uncertainty is reduced and confidence in future reward rises, previously unaligned individuals become aligned simply through acting in their own best interests
2. More robust tracking improves the architecture and allows coordination with decreasing levels of required trust between participants

Eventually, the alignment of parties through the SED process will be robust enough to be applicable to even previously completely unaligned parties who were in direct or violent conflict. At that point, we will be ready to launch Social Environment Design Version Two (SEDv2)

Use the SED system to begin tracking contributions extending beyond the development of system capacities, including all public goods or other actions to help others and the world
1. Normalize for impact and reward using the imputed values of the group as applied to the world at large
2. Drive process improvements in the provisioning of all public goods through the SED process
Once the AI principal agent is confirmed to be fully aligned to the will of the voting public, governance of SED is relinquished by Humanity Unleashed, Inc.
1. Robust method of identity verification, voting/value elicitation deployed prior to governance transfer
2. Bring as many people as possible into the system to ensure that all perspectives are represented and each individual has maximum agency to contribute to collective values and decision making

With the aligned organizations and individuals gathered in the previous phase, use the SED system internally to drive its own development
1. Incorporate AI advances for the principal agent and for enhancements of the reward and tracking systems
Motivate individual action through increasing confidence in the system and likelihood of future reward
1. As uncertainty is reduced and confidence in future reward rises, previously unaligned individuals become aligned simply through acting in their own best interests
2. More robust tracking improves the architecture and allows coordination with decreasing levels of required trust between participants

Use the SED system to begin tracking contributions extending beyond the development of system capacities, including all public goods or other actions to help others and the world
1. Normalize for impact and reward using the imputed values of the group as applied to the world at large
2. Drive process improvements in the provisioning of all public goods through the SED process
Once the AI principal agent is confirmed to be fully aligned to the will of the voting public, governance of SED is relinquished by Humanity Unleashed, Inc.
1. Robust method of identity verification, voting/value elicitation deployed prior to governance transfer
2. Bring as many people as possible into the system to ensure that all perspectives are represented and each individual has maximum agency to contribute to collective values and decision making

With the aligned organizations and individuals gathered in the previous phase, use the SED system internally to drive its own development
1. Incorporate AI advances for the principal agent and for enhancements of the reward and tracking systems
Motivate individual action through increasing confidence in the system and likelihood of future reward
1. As uncertainty is reduced and confidence in future reward rises, previously unaligned individuals become aligned simply through acting in their own best interests
2. More robust tracking improves the architecture and allows coordination with decreasing levels of required trust between participants

Use the SED system to begin tracking contributions extending beyond the development of system capacities, including all public goods or other actions to help others and the world
1. Normalize for impact and reward using the imputed values of the group as applied to the world at large
2. Drive process improvements in the provisioning of all public goods through the SED process
Once the AI principal agent is confirmed to be fully aligned to the will of the voting public, governance of SED is relinquished by Humanity Unleashed, Inc.
1. Robust method of identity verification, voting/value elicitation deployed prior to governance transfer
2. Bring as many people as possible into the system to ensure that all perspectives are represented and each individual has maximum agency to contribute to collective values and decision making