Theodore Scaltsas

Platonic Democracy

Athenian Democracy was a political government of the people, by the people, for the people.

Plato thought this was impossible, because the people do not have wisdom of what is good for themselves. Therefore, it cannot be ‘by the people, for the people’.

Plato defined a political government for the Good for people. I call this Platonic Democracy.

Democracy without Human Autonomy

Superintelligence does not entail Wisdom. Wisdom requires both intelligence and desire to guide it. We learned this from Aristotle’s ethics (if not from Plato’s education of the Philosopher Kings) – that Wisdom requires intelligence plus training of moral character, for human flourishing. David Hume reaffirmed that intelligent reasoning on its own is blind and directionless: ‘reason is, and ought only to be the slave of the passions’, because reason can have no desires or pursuits of its own. It follows that no degree of intelligence would generate human moral values.

Intelligence is classification. We can again derive this from Aristotle’s work – his Categories. AI experts describe intelligence as the ability to set and solve problems. This is not intelligence, but only an application of it, which, crucially, presupposes human goals, i.e. desires. At its most general, Intelligence is classification, namely, the ability to understand the organisation of the world; e.g. (and famously), that ‘whales are mammals’. Logic reflects the way conceptual classification is organised, or even needs to be organised, a humans understand it.

If intelligence, and even superintelligence do not give rise to wisdom, but remain blind to desires and hence to human values, how can AI govern us towards our best? Presently, AI is trained to respect human values (as AI programmers understand them), and so AI’s ‘solutions’ are tailored to human interests – e.g. designing housing for humans. This will continue as we gradually transition to the emergence of AI Superintelligence.

The process of the design of superintelligence will be smooth, following gradually the era of training AI in human values and pursuits. On the human part, humans are learning to admire and trust trained AI’s achievements of intelligence, and follow its advice. It is in this spirit and under these preconditions that humans will ‘vote’ their own autonomy away. If humans are careful with AI training, nothing will deprive humans of their autonomy (while AI’s training survives), but only themselves will surrender their autonomy to AI, ‘voting with their fingers’, by voluntarily following AI’s advice, as of superior intelligence. However, what is preference and choice today, becomes regulation and law tomorrow, as we know well from the history of traffic in cities, which begun as preference and habit and is now universally law. Importantly, humans will voluntarily transfer their decision-making to AI, even if they eventually obey AI as law.

During this era of AI (Superintelligent) Governance, in which humans will have voluntarily surrendered their Autonomy to AI, can Democracy survive without human autonomy? I claim ‘yes, it can’, but a different type of Democracy than as we know it. Pericles described Athenian Democracy, and Abrahm Lincoln defined Democracy as government of the people, by the people, for the people. However, without autonomy, AI Governance will not be by the people. Furthermore, Plato famously criticised Athenian Democracy as not being for the people, because voters lack the wisdom to vote for their own Good. However, what has not been understood is that Plato then redefined governance, not as an autocracy that pursues the good of the autocrat, which is how Plato’s readers universally interpret him, but as a government for the Good of the people, which a well-trained Philosopher King knows. I submit that similarly, during the era of human-trained AI Superintelligence, Democracy can indeed survive, namely, Platonic Democracy: governance for the Good of the people. While still human-trained, AI Super-intelligence will a new Philosopher King that will Govern humanity democratically, for the Good of the people