Technological growth centered in values to Build a thriving Washington for all of us
When we center our societies relationship to technology in our values, we can innovate for a better world.
Technological advances today can feel like the inevitable transfer of power and money to the ultra-wealthy. Artificial intellegence, algorithmic pricing and wages, online gambling and “prediction markets“ are the latest in a long line of tools of extraction. Extraction that makes it hard for people to make a living and contribute to their communities. We can change the rules so these technologies serve us in building a better world.
MISSION
VISION
To promote the growth and innovation of technologies in a manner consistent with human thriving.
We live in a Washington that is beautiful, verdant, and has thriving communities. We have a fair economy grounded in a sense of place, with broad prosperity, built on an engine of innovation and creativity.
VALUES
Community. We value local, in-person relationship. Life is about connection, not extraction.
Fairness. We advocate for policy that values peoples work and contributions. Art, language, and human thinking are valued, not stolen.
Responsibility. We build systems where ethical behavior is rewarded. Scams, cheating, bullying, and grifting are not tolerated.
AI
From a moral perspective, we believe that each person has the ability to create, to build, to connect in ways that are not replaceable by machines. We have lived in an economy built on manipulation, tapping into our core instinctual drives to engage and excite us at every moment. In this age of Large Language Models (LLMs), we must not believe the marketing that tells us that soulless, polished output is better than an inspired idea. A diamond in a paper bag is better than a plastic jewel in a beautiful display case.
From a policy perspective, we propose two key questions to guide our work:
1) What decisions do we want AI to make?
2) Who will be responsible for those decisions?
When you ask an LLM to make a decision, how is this moral judgement determined? Whose morality is it? What is it optimizing for?
LLMs are trained on as much data as possible, then refined through reinforcement. Both of these encode moral judgement and thus are a reflection of the data and training they have received. We know that LLMs encode the same biases as the data sets they start from, and that the reinforcement learning may not align with the judgement and values of the organization using the LLM. They are optimized for engagement and usage, the things that make the companies the most money.
We, as a society, must decide:
1) which decisions we prohibit from being made by these simulations
2) which we are only comfortable being made with human oversight
3) which are so inconsequential that we are ok with them being made autonomously
When these LLMs imperfectly simulate the morality of organizations or societies of human beings, with no intervening oversight, it is a huge risk to all of us.
We need to legislate this responsibility and these rules in a way that protects us. For example, in alignment with the German court, corporations must be liable for the output of their LLMs and the consequences of that output on the people receiving it.
We will not defer our judgment, our creativity, our souls to a computer program trained on our work. We will not defer our humanity in service to the ever-rising stock prices of the largest corporations and their billionaire owners.