Kingson Man, Antonio Damasio (2019), Nature Machine Intelligence 1(10):446–452, Perspective — DOI 10.1038/s42256-019-0103-7
Standing in the record
Filed on 2026-10-01 at the owner’s request for “some Antonio Damasio”.
Unavailable because no full text could be reached, so no NOTE is filed.
What the abstract and the publisher’s page say, to scope it until somebody
reads it:
- It asks whose goals an intelligent machine should pursue, and answers: its own meta-goal of self-preservation, held the way organisms hold it, through homeostasis.
- In organisms, feelings are “a mental expression of the state of life in the body” and regulate behaviour. A machine implementing “a process resembling homeostasis” might acquire “a source of motivation and a new means to evaluate behaviour, akin to that of feelings”.
- The means named are soft robotics and multisensory abstraction. The payoffs claimed are that such machines would “(1) exhibit equivalents to feeling; (2) improve their functionality across a range of environments; and (3) constitute a platform for investigating consciousness, intelligence and the feeling process itself”.
- Fig. 1 is Ashby’s homeostat of 1954, which “exhibited some self-restoring stability”. Fig. 2 compares artificial and natural soft materials.
Is it an instruction for ML practice? By the abstract it is a design
proposal, phrased as “we propose a new class of machines” and “conditions
that would potentially allow machines to care”. It reports no built system
and no measurement. Whether the body turns the proposal into a
recommendation for how to build or train learning systems could not be
checked. It is tagged anthology-candidate because the anthology’s
agents-and-environments topic (systems that perceive, act and are
changed by the consequences) could plausibly hold it. Under ADR-001 that
would make it the anthology’s, so the tag asks for a transfer decision
once it has been read.
Its primary subject is still feeling. Its account of feeling is the one Damasio and Carvalho set out in 2013 (LIT-388), which already frames an action programme “from a bioengineering standpoint”. In the record it sits beside the arguments about whether a machine could feel or be conscious. Seth (LIT-135) holds that consciousness belongs to living, self-maintaining systems and so would need “living AI”. This paper bets that machines with a vulnerable, self-maintained body are the route to feeling. Roberts (LIT-207) argues that affective states need an animate body, which is the premise this proposal tries to build into a machine. It also bears on Bruckner’s welfare criterion of adaptive self-maintenance (LIT-120), which admits machines without requiring sentience. No relation is declared until the paper is read.