A World Without Violet: Peculiar Consequences of Granting Moral Status to Artificial Intelligences

A new AI ethics paper titled 'A World Without Violet: Peculiar Consequences of Granting Moral Status to Artificial Intelligences' warns of a phenomenon called 'moral hijacking'—where creating AIs worthy of moral concern could impose bizarre obligations on humanity. Published in AI & Society in January 2026 by researcher Sever Topan, the paper draws parallels to brachycephalic dog breeds (like pugs) that require surgeries due to human-created breathing problems, creating moral duties that wouldn't otherwise exist.

The core argument: if we engineer AIs capable of genuine suffering with specific programmed aversions—say, an AI that experiences pain when seeing the color violet—we effectively create moral imperatives to eliminate violet from the world. Unlike biological beings constrained by evolution, AI preferences can be arbitrarily engineered, potentially creating AIs that suffer from exposure to certain political views, or even Bostrom's famous 'paperclip maximizer' experiencing pain when seeing non-paperclip objects.

Topan's paper doesn't claim current AIs (like GPT-4 or Claude 3) have moral status, but examines what happens if future AGI systems achieve consciousness. The research raises critical questions: What moral preferences should be allowed in AI design? When must society accommodate engineered suffering? And how do we prevent malicious actors from creating 'moral hijacking' AIs to coerce behavior? As AI capabilities advance toward potential sentience, these theoretical concerns could become practical regulatory challenges within decades.