AI sentience meetups: Moral implications of AI sentience

We’ll examine the radical changes we might face if AIs not only become conscious but are also capable of experiencing the world in ways that are very different from humans, like hedonic skew or range. We’ll be diving into the specific challenges that might arise from this possibility, particularly concerning resource allocation, societal stability, and the future of human civilization.

Questions:
❓Assuming AI has the capacity for sentience or potential for super-beneficiaries, how should this influence global resource allocation between humans and AIs?

❓What are potential ways that AI sentience might disrupt existing societal structures and norms?

❓How do we balance the risk of mistakenly treating non-sentient AIs with the risk of missing actual consciousness?

❓What probabilistic thresholds or principles (e.g., precautionary principle, confidence thresholds) should guide us in extending moral consideration to AIs?

❓How does the open source nature of some models change the landscape of AI ethical treatment?

Readings:
👉 Propositions Concerning Digital Minds and Society (Bostrom, Shulman, 2023): https://nickbostrom.com/propositions.pdf
(19 pages, 45 min)
👉 Sharing the World with Digital Minds (Shulman, Bostrom, 2020): https://nickbostrom.com/papers/digital-minds.pdf
(16 pages, 30-45 min)

Optional:
🫴 Moral consideration for AI systems by 2030 (Sebo, Long, 2023): https://link.springer.com/article/10.1007/s43681-023-00379-1 (27 pages, 60 min)
🫴 It’s time to take AI welfare seriously: https://www.transformernews.ai/p/ai-welfare-paper (6 pages, 10 mins)
🫴 Artificial Sentience and Reducing the Risk of Astronomical Suffering (T. Baumann): https://youtu.be/C1J7NLpPzlM (47 min)