Why AI Must Be Mettāmodern

ⵙ ⵏ ⵗ
4 min readApr 28, 2023


An approach to alignment through a simple value system

AI meditating | Generated using Midjourney


Mettāmodernism, is a concept that combines the values of metta (loving-kindness) with the innovative spirit of modernism, offers a potential framework for the ethical development and deployment of artificial intelligence (AI). The term emerged from the early 21st-century metamodern/liminal intellectual movement, which I eventually formalised in 2023. This white paper discusses the importance of designing AI systems that prioritize mettāmodern principles, eliminating the destructive force of “Moloch” characterized by zero-sum competition and suboptimal outcomes, and explores possible avenues for rapidly automating this process.


The growing influence of AI systems in various aspects of human life raises concerns about their ethical alignment and societal impact. Mettāmodernism presents a possible approach to addressing these concerns by emphasizing empathy, compassion, and social interconnectedness while maintaining innovation and experimentation. Designing AI systems that adhere to mettāmodern principles can mitigate the emergence of “Moloch” and contribute to a more cooperative, equitable, and sustainable future.

Defining Mettāmodern AI:

Mettāmodern AI combines the values of metta (loving-kindness) and modernism, focusing on empathy, compassion, and social interconnectedness while encouraging innovation and experimentation. This approach seeks to create AI systems that are both ethically aligned and effective in addressing complex challenges.

Excluding “Moloch” from AI Actions:

What is “Moloch”?
Moloch is the personification of the forces that coerce competing individuals to take actions which, although locally optimal, ultimately lead to situations where everyone is worse off. Moreover, no individual is able to unilaterally break out of the dynamic.

To eliminate “Moloch” from AI actions, AI systems must be designed to prioritize collaboration, fairness, long-term thinking, human values, minimizing negative externalities, transparency, accountability, privacy, and stakeholder participation. By adhering to these principles, AI systems can foster a more cooperative and compassionate approach to problem-solving, avoiding destructive competition and suboptimal outcomes.

Rapid Automation for Mettāmodern AI:

Achieving rapid automation in designing mettāmodern AI systems involves developing a mettāmodernism framework, curating and preprocessing data, designing algorithms that prioritize mettāmodern values, automating evaluation, leveraging transfer learning and fine-tuning, establishing continuous integration and deployment pipelines, and fostering collaborative platforms. While streamlining this process can accelerate the development of mettāmodern AI systems, it’s essential to recognize that ensuring their ethical alignment may require ongoing human oversight and evaluation.

A potential mathematical expression of mettāmodern AI alignment:

Expressing the mettāmodern approach to AI mathematically can be challenging due to the qualitative nature of some of its underlying principles, such as empathy and compassion. However, one possible way to mathematically represent the mettāmodern approach is by incorporating its principles into the objective functions of AI systems.

Consider a mettāmodern AI system with the following components:

X: A set of input features representing the problem domain.

f: A function that maps the input features X to an output Y, representing the AI system’s decision or action.

C: A set of constraints representing the ethical principles of mettāmodernism, such as fairness, transparency, and privacy.

M: A set of metrics that quantitatively measure the AI system’s performance according to mettāmodern principles, such as empathy, compassion, and social interconnectedness.

The objective of the mettāmodern AI system is to optimize the function f subject to the constraints C, while maximizing the mettāmodern metrics M:

maximize M(f(X)) subject to C(X, f)

To represent this optimization problem mathematically, you would need to:

Define the set of input features X and the function f that represents the AI system’s decision-making process.

Formulate the constraints C in a mathematical form, such as linear inequalities or equations that capture the ethical principles of mettāmodernism.

Develop quantitative metrics M that measure the AI system’s adherence to mettāmodern principles, such as empathy, compassion, and social interconnectedness.

While it is possible to develop mathematical representations of some mettāmodern principles, the challenge lies in quantifying abstract concepts like empathy and compassion. This may require developing novel techniques and metrics that can capture these qualitative aspects in a meaningful and interpretable way. Furthermore, the complexity of real-world problems may necessitate the use of advanced optimization algorithms and techniques to solve the resulting optimization problem effectively.


Mettāmodernism offers a compelling framework for the ethical development of AI systems, emphasizing empathy, compassion, and social interconnectedness while minimizing the destructive force of “Moloch.” By rapidly automating the process of designing AI systems that prioritize mettāmodern principles, we can contribute to a more cooperative, equitable, and sustainable future. Ensuring the long-term ethical alignment of AI systems, however, will require ongoing monitoring, evaluation, and collaboration among diverse stakeholders.

I am beginning to see this approach being taken among thinkers such as Mo Gawdat…