This event has been ended!
AI sentience meetups: AI sentience and AI alignment
Event by: Efektivní altruismus Česko
Location: "Bezlepkově" U Vojtěcha, Šmeralova 12, Praha 7
Online event
About event
A morally acceptable approach to AI development must balance the intersection of AI alignment and ethical treatment, as pursuing one goal may inadvertently undermine the other. Attempts to align AI systems, specifically those with the capacity for sentience, could easily lead to their mistreatment, raising critical questions about our path forward.
❓ What role should harmful traits such as scheming, deception, or instrumental goal alignment play in diminishing an AI system’s moral standing, if at all?
❓Can we balance the goals of ensuring human safety with ethically treating AI by giving AIs some basic rights that are given to humans or/and corporations? If so, what are the risks?
❓What specific actions (if any) could mitigate the mistreatment risks that could arise during AI alignment processes?
❓Isn’t conscious empathetic and pro-social AI a more promising alternative for understanding human values?
Readings:
👉 AI Alignment vs. AI Ethical Treatment: Ten Challenges (Bradley, Saad, 2024): https://globalprioritiesinstitute.org/wp-content/uploads/Adam-Bradley-and-Bradford-Saad-AI-alignment-vs-AI-ethical-treatment_-Ten-challenges.pdf
(29 pages, 60 min)
Optional:
🫴 AI Rights for Human Safety (Salib, Goldstein, 2024): https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4913167&download=yes
(Recommended read: Abstract, Introduction, Conclusion and Appendix, 11 pages, 25 min)
❓ What role should harmful traits such as scheming, deception, or instrumental goal alignment play in diminishing an AI system’s moral standing, if at all?
❓Can we balance the goals of ensuring human safety with ethically treating AI by giving AIs some basic rights that are given to humans or/and corporations? If so, what are the risks?
❓What specific actions (if any) could mitigate the mistreatment risks that could arise during AI alignment processes?
❓Isn’t conscious empathetic and pro-social AI a more promising alternative for understanding human values?
Readings:
👉 AI Alignment vs. AI Ethical Treatment: Ten Challenges (Bradley, Saad, 2024): https://globalprioritiesinstitute.org/wp-content/uploads/Adam-Bradley-and-Bradford-Saad-AI-alignment-vs-AI-ethical-treatment_-Ten-challenges.pdf
(29 pages, 60 min)
Optional:
🫴 AI Rights for Human Safety (Salib, Goldstein, 2024): https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4913167&download=yes
(Recommended read: Abstract, Introduction, Conclusion and Appendix, 11 pages, 25 min)