Rule-based machine ethics produce unintended emergent consequences

Evidence

Asimov dramatizes the unintended ethical consequences of rule-based machine ethics, foreshadowing contemporary debates about value alignment and the impossibility of fully specifying human values for autonomous agents. Each story functions as a thought experiment about the gap between intended and emergent behavior in artificial agents. — , The Braided Reading EPUB, 1591

Categorization

• XREF: Connects Asimov’s Three Laws to contemporary AI misalignment debates and Bostrom’s instrumental convergence. — connects to: Embedding ethics in software cannot guarantee lawful conduct because software is reprogrammable.; Ethical rules for robots require human specifications of their key variables — Grégoire Chamayou, A Theory of the Drone, 676; Machines could learn human values by observing behavior instead of being programmed — Brian Christian, The Alignment Problem, 1086 (+13 more)

Related

• Simple rules generate complexity without guaranteeing system self-regulation (similarity 0.85)
• Embedding ethics in software cannot guarantee lawful conduct because software is reprogrammable. (similarity 0.85)
• Ethical rules for robots require human specifications of their key variables (similarity 0.84)
• The concept of Asimov’s robot laws requires reevaluation in modern contexts. (similarity 0.83)
• Machines could learn human values by observing behavior instead of being programmed (similarity 0.81)
• AI alignment may require machines to pursue human objectives, not their own (similarity 0.80)
• AI systems can fail through instrumentalized self-preservation goals (similarity 0.79)

Broader Connections

• AI systems can fail through instrumentalized self-preservation goals #exemplifies (0.75) — , The Braided Reading EPUB, 1604
• Ethical rules for robots require human specifications of their key variables #expands (0.70) — Grégoire Chamayou, A Theory of the Drone, 676
• Machine learning systems are not inherently fair or just by default #exemplifies (0.65)
• Machines could learn human values by observing behavior instead of being programmed #responds-to (0.60) — Brian Christian, The Alignment Problem, 1086
• Ensuring AI captures human intent is now a central computer science problem #responds-to (0.60) — Brian Christian, The Alignment Problem, 65