Governance

• OpenAI was intended as a non-profit to keep AI aligned to public interest. It is currently a non-open for-profit, with virtually all checks-and-balances stripped away.
• edX was advertised as a consortium designed to keep MOOCs operating in the public interest. It generated $800M to the founding institutions, sold to a for-profit, and crashed-and-burned.
• Elite institutions, like MIT, Harvard, and Stanford, have widespread research fraud, as behavior converges to incentives. This hurts students as schools invest in things which can’t work.
• Communism seemed like a good idea, until folks like Stalin took power.

Many initiatives continue to function well so long as the founder is still there, but decline as soon as the founder leaves. A few areas to think about:

• Lessig sets up one framework in Code and Other Laws of Cyberspace, which emphasizes that how we set up technology is one form of governance (among four: law, norms, market, and architecture).
• This technocratic perspective is balanced by the game-theoretic approach in de Mesquita’s The Dictator’s Handbook: Why Bad Behavior is Almost Always Good Politics, which outlines how, in the long run, governance converges to incentives and markets.

There have been many successful attempts to set up open ecosystems, including the GPL ecosystem built by the FSF and, to a lesser extent, the Creative Commons ecosystem. Ideas can come from government (what makes democracies work? what makes it fail?), business, and non-profits.

As the other groups discuss what we want for privacy, family rights, open science, etc., think through of what we would need to structure an organization which can maintain this sort of alignment for the next hundred years – in other words, beyond the lifetimes of the founding members. Think through:

• Legal agreements
• Transparency models
• Governance
• Licenses
• Consortium models

How do we set up a legal and technological framework to keep us honest and operating in the public good?