![Cover Image for Analytic Philosophy After Dark: Moral Realism and the AI Safety Debate [public]](https://images.lumacdn.com/cdn-cgi/image/format=auto,fit=cover,dpr=2,background=white,quality=75,width=400,height=400/event-covers/fw/76f2cdcd-5f98-4113-9692-52a87834c9c5)
Analytic Philosophy After Dark: Moral Realism and the AI Safety Debate [public]
Join us at one of our favorite bars to discuss the connection between metaethics and AI risk.
It may seem that the question of how dangerous AI is and the question of whether there are any moral truths (metaethics) couldn’t be any further apart. But an influential strand of AI pessimism seeks to argue for what is called the Orthogonality Thesis:
There are no objective moral truths (moral anti-realism).
If there are no objective moral truths, there is no reason to think that the smarter an AI gets the more likely it will have values that align with humanity’s.
Since there is no reason to think that superintelligent AI will have values that are aligned with humanity’s, superintelligent AI is a threat to humans.
What do you think of this argument? Come discuss with us.
Possible discussion questions
Is the Orthogonality Thesis true? Is there a connection between the objectivity of moral truths, and AI risk?
Are there moral facts?
What is the metaphysical status of moral claims? Are they objective, relative, or fictional?
If moral realism is true, would there be less reason to worry about AI safety?
Optional reading
Subscribe to our Luma calendar: https://lu.ma/apad