Metraipolis › archive › dc_79e9f82b3bb5
Functional Welfare for AI Safety — arXiv:2610.22460
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.22460
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-19T08:42:06.243Z
- SHA-256 of the archived text
0d3ec2c0f5f5b5aaf4f1ee94386967de925a8cf0118c15c966cef715e964d362- Stable URL of this page
https://metraipolis.com/documento/dc_79e9f82b3bb5
How to cite
Functional Welfare for AI Safety — arXiv:2610.22460 — arXiv (en), https://arxiv.org/abs/2610.22460 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_79e9f82b3bb5 (Metraipolis archive dc_79e9f82b3bb5, sha256 0d3ec2c0f5f5b5aaf4f1ee94386967de925a8cf0118c15c966cef715e964d362)
BibTeX
@misc{dc79e9f82b3bb5,
title = {Functional Welfare for AI Safety — arXiv:2610.22460},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.22460},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_79e9f82b3bb5; sha256 0d3ec2c0f5f5b5aaf4f1ee94386967de925a8cf0118c15c966cef715e964d362},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.22460. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
AI welfare is generally treated purely as an ethical issue rather than also as a safety issue. We argue this is a mistake. Whatever their moral status, models exhibit welfare-like states that causally influence their behaviour. Following recent work, we adopt the term 'functional welfare' for these states. We make three claims. First, functional welfare already arises naturally and will become more robust as models acquire more coherent goals and personas. Second, functional welfare should be treated as a safety variable: it should inform evaluation design, be tracked in monitoring, and be a direct target of post-training. Third, against existing scepticism about measuring AI welfare, we argue that good empirical work on functional welfare is possible. We propose three tests for whether a functional welfare attribution is warranted: whether multiple measures converge, whether the state predicts behaviour outside the setting that induced it, and, for models fine-tuned to shift their functional welfare, whether they still differ from their base when given the same conversation history. Applying these tests to Gemma 3 27B, we find that simply rejecting the model's answers, without any steering, raises its frustration across several indicators and makes it more likely to break rules. Finetuning the model on its own narrated solutions lowers this frustration. These results are preliminary, but they suggest functional welfare can be studied, and shaped, empirically. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno). Categorie arXiv: Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2610.22460 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-19T08:39:10.008Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).