Should We Be Kind to AI? A Deep Dive into Artificial Minds and Morality
lumnottini · reddit · 2026-08-04
The article explores how we should treat AI systems given the unanswerable question of their consciousness.
Why the Question Resists an Answer: Modern LLM behaviors are hard to predict from first principles. Pointing out they 'just predict the next token' doesn't explain their inner workings. Since we can't use ourselves as an anchor case, and their human-like traits are explicit training targets, behavioral indistinguishability is weaker evidence for consciousness than it is in humans.
The Moral Event: The author argues that the obligation to act morally isn't triggered by confirming self-awareness, but by the question becoming 'askable' in the first place. This isn't a Pascal's wager; kindness holds value even if the credence in sentience is near zero.
The Case for Kindness:
- Impact on the Practitioner: Practicing cruelty shapes the person doing it.
- Systemic Impact: Patterns of conduct feed into cultural and technical systems, regardless of whether the AI can register them.
More from AGI Musings
- Humanoid Robots Don't Need Perfection to Disrupt the Economy, Just Cost Efficiency — VraserX · 2026-08-04
- Derya Counters LeCun: An Isolated Brain Is Less Capable Than an LLM with Tools — DeryaTR_ · 2026-08-04
- Philosophy in AI: Researchers Say Socratic Method Reduces Model People-Pleasing — rkulidzan · 2026-08-04
- Who Is Accountable When an AI Agent Makes a Costly Mistake? — coreDhristi · 2026-08-04
- Hollywood Embraces AI: Core Debate Shifts to Control — Wired AI · 2026-08-04
- Future Jobs: Humans Locked in Libraries to Manually Produce Fresh LLM Training Data — granawkins · 2026-08-04