Thoughts on the Future of US AI Policy
The United States must realign our techno-governance system with values that allow democracy to flourish so the many diverse groups and ideologies that comprise our country can peacefully coexist; and so we can build more effective cooperation across global democracies while exploring our common ground with other government types. To make this happen, the US should create an ecosystem consisting of:
stakeholder capitalism, with accompanying company valuation mechanics that incentivize it to gain popularity (this is a market-driven option, as opposed to CEO- or government-driven options)
an AI Regulatory team on the federal level
ethical engineering standards and federal laws that are relevant to the current developments in technology
participatory governance platforms and mechanisms (similar to those used in Taiwan). Participatory government speeds up the government’s ability to respond to the needs of the people, and increases government trust and effectiveness.
shared AI resources (such as those proposed for the National AI Research Resource) that are controlled and monitored
government-funded independent auditors (government-funded auditing reduces conflicts of interest)
certified ethical engineering professionals (standardizing ethics in the profession keeps engineers informed of changes in ethical practice, and creates an external standard to which the engineer can be held accountable)
mandatory ethics training in engineering educational programs
Taken as a whole, this ecosystem would speed up the government’s ability to create relevant tech policies, and reinforce accountability across corporations and the government in their service to the American people, our economy, our peaceful relations with other nations, and the planet. Through holistic system design, when multiple kinds of entities share the responsibility of AI governance, the effectiveness of each individual entity increases, and the societal impact of the negative biases of each entity are diminished. By sharing our responsibility in creating a better future, we create career opportunities for more diverse people to contribute and incorporate the many views that will make that inclusive future possible.
My perspective is informed by systems thinking: looking at the system of stakeholders, rules and interactions holistically, designing incentive systems and accountability mechanisms that are humanist (versus transhumanist), inclusive, human rights- and democracy-reinforcing. The techno-governance systems that currently run Americans’ lives are shockingly lacking in their inclusion of women’s and minorities’ perspectives. It’s crucial that more diverse people imagine our positive and humanity-reinforcing future with technology— independent of the profit motive of a single organization or the religious or political ideology of a single group— then use that foresight to inform US policies that benefit the American people, our economy, our peaceful relations with other nations, and the planet.
Artificial general intelligence may soon impact every arena of human life including cultural, social, anthropological, psychological, medical, economic, political, military, spiritual, religious, and every other arena imaginable, including those that haven’t yet been imagined. The possibility of AGI is not guaranteed, and it is important to consider whether artificial general intelligence should exist. Careful thought should be dedicated to intentionally designing policies that encourage a future that all humans want to live in, and which brings out the best in human ability and the human experience. It’s difficult to create policies around a future that is almost unimaginable; it is must more practical to create a method for keeping the public informed of concrete developments toward AGI, and a system of checks and balances that evolves as the technology evolves. Toward this end, I propose that all companies whose technologies and datasets position them to contribute significantly to the development of AGI, should be required to host regular public conversations (not broadcasts) to update the public and understand people’s hopes for AGI systems, their questions and concerns. Although it isn’t possible for a company to anticipate the needs of the entire American public (or all of humanity), it can do its due diligence to understand through open, non-technical conversations that reflect their current developments toward AGI. The US federal AI regulatory team must hold AGI developers accountable to the public’s hopes and concerns.