Luciano Floridi (2023), Philosophy & Technology 36(1): 15 (editorial) — DOI-10.1007/s13347-023-00621-y
Key takeaways
- LLMs are claimed to have no understanding, thought or reasoning: they handle “the formal structure, and not … the meaning of the texts they process” (pp. 1–2). The support is a list of known failure modes and one screenshot (Fig. 2), not an argument.
- “Stochastic parrots” is judged only partly apt, because LLMs synthesise rather than repeat (p. 3). Floridi’s preferred images are search-engine autocomplete, the lazy student’s eclectic essay, and Manzoni’s trickster who swallows tow and pulls out endless “tape” (pp. 3–4).
- The thesis comes in the last three paragraphs (pp. 5–6). LLMs are “a form of agency never seen before, because it is successful and can ‘learn’ and improve its behaviour without having to be intelligent to do so”. What separates this from a sea wave, which “makes the difference, but can make nothing but that difference”, is the capacity to learn to make “a different or better difference”. “We have liberated agency from intelligence.”
- The agency notion is attributed by citation to Floridi & Sanders 2004 (LIT-898), in one parenthesis (p. 6). None of that paper’s apparatus (levels of abstraction, its criteria for agency) appears here.
Standing in the record
Filed on 2026-10-10 at the owner’s request, as part of the line of work behind Floridi’s “AI as Agency without Intelligence: On Artificial Intelligence as a New Form of Artificial Agency and the Multiple Realisability of Agency Thesis”, Philosophy & Technology 38:30 (2025), DOI-10.1007/s13347-025-00858-9, filed as LIT-897. That paper takes this editorial’s title and develops its thesis. It was read later the same day (NOTE-702), and nothing in this entry depends on what it says. The anthology holds neither work. This one carries no instruction for machine-learning practice, so it is held here.
The reading places it Active, but as a source for the slogan and its first framing, not as an argument for it. Its value to the record’s agency line is that it states a sharp position: success at acting, plus learning to improve, is enough for agency, and no intelligence or understanding is needed. That puts it at the permissive end against the record’s organizational account of minimal agency (LIT-566). It does not give a criterion that would answer QUESTION-009.
No extends: to Floridi & Sanders 2004 (LIT-898) is written. The editorial would stand without that paper: it cites it once, for the phrase “new form of artificial agency”, and uses none of its machinery. The link is a citation, not a dependence.