Don’t blame AI for the Iran school bombing | AI (artificial intelligence)

https://www.theguardian.com/technology/2026/apr/01/dont-blame-ai-for-the-iran-school-bombing

Publish Date: 2026-04-01 13:47:00

Certainly! Here’s a concise summary of the points made by Anthony Lawton and Dr. Felicity Mellor within the article referenced:

Dislocation of Human Responsibility: The article argues against attributing mistakes to AI systems, highlighting that such attributions remove the human designers, authorizers, and executors from their moral and legal responsibilities.
Language and Accountability: It stresses the importance of clear language in assigning accountability. Obscuring human responsibility, even through subtle shifts in terminology, undermines moral accountability and public scrutiny.
Civic Responsibility Over Technical Faults: By speaking of “AI errors” instead of “human errors,” the language diverts focus from the human chain of command to technological systems, compromising civic responsibility and oversight.
Conflating Humans and Machines: Anthropomorphic language used to describe AI behaviors – such as “conniving” or “cheating” – wrongly invests AI with agency and moral choices, disregarding where actual responsibility lies.
Moral Agency Attribution: Accurately attributing moral agency where it belongs — to humans behind the technology — is crucial for accountability, especially in scenarios where AI causes harm.

This summary captures the essence of the arguments surrounding the ethical responsibility in AI accountability and the need for precise, clear public discourse on the topic. If interested in further discussion or to share your thoughts on this subject, consider sending your letter to the Guardian for possible publication in their letters section.