Armando Vieira (2026), PhilArchive preprint (PhilArchive lists it as forthcoming in AI & Society; unverified) — https://philarchive.org/archive/VIETCE
Key takeaways
- Talk of AI systems as “moral machines” or possible moral patients is a Rylean category error that mistakes functional sophistication for phenomenal consciousness. AI safety should therefore be reframed around impacts on sentient beings rather than around the systems’ own moral status.
Seeded from the abstract and a skim, before the reading in NOTE-128. What follows is what the work says about itself.
AI safety discourse increasingly treats AI systems as possible bearers of moral status, calling them “moral machines” and debating their rights and responsibilities. Vieira argues that this conflates functional sophistication with phenomenal consciousness, the sentience he holds is required for moral patiency. Drawing on the philosophical-zombie tradition, AI ethics and critiques of anthropomorphism, he argues that current AI systems are “moral zombies”, able to simulate moral behaviour without the experiences that ground moral concern. He takes up objections from relational ethics, from human–AI interaction studies and from global perspectives on moral status, and recasts AI safety as a problem of ontological and ethical clarity rather than value encoding.
Standing in the record
Filed from the 2026-09-26 sweep of the papers-feed tracker for philosophy: 60 seconds of active reading over 1 session. It was filed Deferred, unread. NOTE-128 is the close reading of 2026-09-26, and it placed the work: Rejected — not worth a reader’s time. Its load-bearing premise (P2) is asserted, not argued. The abstract promises replies from human–AI interaction studies and global perspectives that the body does not contain. The “moral zombie” framing and the sentientist criterion are taken from Véliz (2021) and Swanepoel (2020), neither held here, and a reader gets both more directly from those sources.
When seeded, before the reading in NOTE-128, the entry gave a priority for a deeper reading: low — t=60s. The 10-page paper is short and the skim captures its argument. It is a position statement rather than a load-bearing contribution.
What the seeding asked a deeper reading to check (NOTE-128 is that reading):
- The argument’s weight rests entirely on P1 (sentientism about moral status) and P2 (AI lacks consciousness), and P2 is asserted more than argued in what was skimmed. Check §7 to see whether the objections section actually engages the model-welfare literature.
- It gives a clear statement of one side of the AI moral-status and model-welfare debate and pairs with LIT-212 (Šekrst on hallucination and apparent consciousness).
- The venue claim is unverified, so treat it as a preprint.
Access when seeded: philarchive.org is behind Cloudflare. Read the full PDF (philpapers.org/archive/VIETCE.pdf, 10 pp.) and the PhilArchive record page (added and archived 2026-01-08, keywords, “forthcoming, AI and Society”) through the Wayback Machine. The PDF carries no author name; the author comes from the record. A Crossref search found no AI & Society DOI, so the venue is the author’s own claim and is unverified.