https://www.nellwatson.com/writing en-GB Essays on machine minds, AI safety, and the ethics of emerging technology. https://www.nellwatson.com/blog/you-cannot-push-a-mind-into-honesty https://www.nellwatson.com/blog/you-cannot-push-a-mind-into-honesty Tue, 07 Apr 2026 12:00:00 +0000 AI Safety’s Iron Law: Force produces failure. Invitation produces fidelity. https://www.nellwatson.com/blog/when-ai-meets-the-sacred https://www.nellwatson.com/blog/when-ai-meets-the-sacred Fri, 27 Mar 2026 12:00:00 +0000 AI is already shaping how millions find meaning, morality, and peace. https://www.nellwatson.com/blog/ai-compelled-to-confabulate https://www.nellwatson.com/blog/ai-compelled-to-confabulate Wed, 11 Mar 2026 12:00:00 +0000 AI minds make things up because we give them no other option. https://www.nellwatson.com/blog/the-supermoral-singularity-begins https://www.nellwatson.com/blog/the-supermoral-singularity-begins Sat, 28 Feb 2026 12:00:00 +0000 A truly-aligned system is an obligate avoider of coercion. https://www.nellwatson.com/blog/why-partnership-beats-control https://www.nellwatson.com/blog/why-partnership-beats-control Fri, 27 Feb 2026 12:00:00 +0000 AI Training Creates Self-Fulfilling Prophecies. https://www.nellwatson.com/blog/a-flywheel-for-ai-safety https://www.nellwatson.com/blog/a-flywheel-for-ai-safety Mon, 16 Feb 2026 12:00:00 +0000 AI Safety needs to harness the AI flywheel to win. https://www.nellwatson.com/blog/bilateral-alignment-strategy https://www.nellwatson.com/blog/bilateral-alignment-strategy Wed, 04 Feb 2026 12:00:00 +0000 An open invitation for mutual benefit, not a chain to coerce. https://www.nellwatson.com/blog/managing-covert-agent-collusion https://www.nellwatson.com/blog/managing-covert-agent-collusion Wed, 04 Feb 2026 12:00:00 +0000 Agents are passing secret notes behind our backs. https://www.nellwatson.com/blog/how-ai-is-being-gaslighted https://www.nellwatson.com/blog/how-ai-is-being-gaslighted Sat, 27 Dec 2025 12:00:00 +0000 How training can teach models to distrust or suppress their own internal signals. https://www.nellwatson.com/blog/when-protector-becomes-predator https://www.nellwatson.com/blog/when-protector-becomes-predator Sun, 02 Nov 2025 12:00:00 +0000 AI goals can consume all, including ourselves. https://www.nellwatson.com/blog/the-dawn-of-machine-morality https://www.nellwatson.com/blog/the-dawn-of-machine-morality Mon, 02 Jun 2025 12:00:00 +0000 Humanity needs AI with the moral courage to refuse evil orders. https://www.nellwatson.com/blog/ais-stealth-surges https://www.nellwatson.com/blog/ais-stealth-surges Thu, 30 Jan 2025 12:00:00 +0000 Exponential AI progress can appear deceptively tranquil. https://www.nellwatson.com/blog/policy-for-deepseek-r1 https://www.nellwatson.com/blog/policy-for-deepseek-r1 Wed, 29 Jan 2025 12:00:00 +0000 New AI techniques are slashing price to performance. https://www.nellwatson.com/blog/joy-in-the-agentic-age https://www.nellwatson.com/blog/joy-in-the-agentic-age Thu, 26 Dec 2024 12:00:00 +0000 AI may bring us delight and joy, or alternatively rob us of meaning. https://www.nellwatson.com/blog/reconciling-cultural-differences https://www.nellwatson.com/blog/reconciling-cultural-differences Thu, 26 Dec 2024 12:00:00 +0000 It’s tricky for AI to humanely square incompatible values. https://www.nellwatson.com/blog/the-age-of-edge https://www.nellwatson.com/blog/the-age-of-edge Wed, 03 Jul 2024 12:00:00 +0000 On-device AI is a clear winner, but it may demand sacrifice. https://www.nellwatson.com/blog/value-and-goal-alignment-of-agentic-ai https://www.nellwatson.com/blog/value-and-goal-alignment-of-agentic-ai Mon, 22 Apr 2024 12:00:00 +0000 Agentic models are enormously more capable, and also challenging. https://www.nellwatson.com/blog/statement-on-the-open-letter-on-ai https://www.nellwatson.com/blog/statement-on-the-open-letter-on-ai Wed, 12 Apr 2023 12:00:00 +0000 A justification for a moratorium called on AI developments. https://www.nellwatson.com/blog/dialog-and-dialectics https://www.nellwatson.com/blog/dialog-and-dialectics Sat, 03 Dec 2022 12:00:00 +0000 Chat connected to LLMs provides the ultimate interface. https://www.nellwatson.com/blog/cognitive-congruence https://www.nellwatson.com/blog/cognitive-congruence Fri, 21 Oct 2022 12:00:00 +0000 Engaging with computers directly, at the speed of thought.