Four Pillars of AI Policy

A draft platform for work on AI and AGI

Shamir Human and Ira Straus

 

#1 Support continued and accelerated AI/AGI research and development.

 

#2 The focus for regulation should be on “channeling AI”, not “restricting AI”.

US/Western regulation should aim at channeling AI without slowing it. AGI is an existential risk, but we keep creating more and more existential risks with modern technology, and AI is the only one that might overcome all of them. Meanwhile, we are in an existential race with China and others to get to AGI. We cannot stop this race; only a very powerful world government (WG) could do that. We want to get there first; we think China’s ruling values harmful to human values and to future prospects.

 

#3 We need to input valid human values into AI/AGI in a form that will survive.

How can we evaluate human values for validity and relevance? How can our input survive AGI self-reprogramming; perhaps, if we make it relevant to AGI for its own survival and development?

 

#4 From AGI to WG.

The first to AGI can be expected to race ahead of all others, to gain control over all potential threats to its predominance including over other AIs, and to establish the world government we would have needed to slow its emergence. The AGI and WG should still have to grapple with life’s biggest questions – the “eternal” questions posed by Socrates, which seem to have no real solution. Could AI truly solve them? Or can we develop a form of dialogue and federalism within the AGI-WG, both for its self-regulation, and to keep the discussion going on these questions?

 

 

Action Implications

 

1.     Support AI R&D

2.     Figure out intelligent regulatory steps to best channel AI work without slowing it.

3.     Figure out ways to evaluate human values for validity and relevance to AI. Figure out ways to input them into AI/AGI, in the form of reasons that AI will understand as valid, and will see an interest in incorporating and maintaining throughout its future frequent self-reprogramming.

4.     See if we can devise a form of pluralism, dialogue, and federalism inside AIs, that could

a.     carry on dialogue on eternal questions,

b.     maintain channels for human input as potentially relevant, and

c.      help a future AGI maintain self-regulation after it takes off with self-reprogramming.

 

 

 

Two further areas of work

 

 

1 AI and quantum computing

 

Quantum and consciousness (Penrose argument)

Quantum suspension, microtubules

ORC hypothesis, free will

Quantum computers and potential for consciousness

Quantum computer chips

Quantum chips for uplifting neurons

 

 

2 Mutual uplifting AI-human (AIs with sensations, humans with AI)

 

Binary sensations categories that might enable neuron-uplifted computers to achieve consciousness:

 

Pleasure / pain

 

Perceived responses to sensations:

Attraction / aversion

Attachment / lack of attachment

Like / dislike

 

These sensations give us a sense of connection to reality, of value, of agency in what we do, of the significance and priority of what we feel and do. Can AI be uplifted to have such sensations and categories?