GoKawiilNew 'decision model' AI systems tested on classic Trolley Problem ethics dilemma
A developer ran the Trolley Problem thought experiment through Jev and twelve other 'decision models,' including OpenAI's Luna and eleven other local models, plus a baseline called Laya. These decision models are a new category of AI that output discrete choices with confidence probabilities rather than text, and are marketed as much faster and cheaper than large language models while retaining comparable reasoning ability. The experiment repeated the classic ethical dilemma in multiple phrasings across all models to compare their responses.
GoKawiil's interpretation of the reporting above, not reported fact.
The test suggests that as decision models get adopted for fast, cheap automated choice-making, their handling of ambiguous ethical tradeoffs could matter for real-world deployments where such systems make consequential calls. Comparing multiple vendors' models on the same moral dilemma offers an early, informal benchmark for how consistent or divergent these systems' 'values' are. This kind of probing could become more important as decision models move from novelty into production use for automated decision-making.
- A new class of AI called 'decision models' outputs choices with probabilities instead of text, and claims to be far cheaper and faster than LLMs like Claude.
- A writer tested Jev, OpenAI's Luna, eleven other decision models, and baseline model Laya on the classic Trolley Problem.
- The experiment repeated the dilemma in multiple phrasings to compare how different decision models respond to the same ethical tradeoff.
Source: zachlandes.com — Zach Landes, 2026-10-15
Published there as: “Jev Meets the Trolley Problem”
Read the original report → The summary and analysis above are GoKawiil's own, written from reporting by the source above. Facts and quotes belong to the original publisher.