{"CACHEDAT":"2026-05-13 09:33:16","TRANSLATEDAT":"2026-05-13 09:33:16","SOURCESIGNATURE":"870824410e5d47f07c74761410638bd1dfca965fe0dd439c19b79b771c0f2725","SLUG":"glossary-ma8b2MAZra","MARKDOWN":"# Algorithms\n\nAn algorithm is a sequence of instructions, automatically executed by a computer (UNESCO, 2017). It input into output by performing calculations, sorting or classifying data, making decisions, or automating repetitive tasks. Algorithms range from very simple (e.g., a fixed rule for sorting a list) to highly complex (e.g., the recommendation system of a social media platform).\n\nAlgorithms are designed by humans (Council of Europe, 2020). Even in modern machine learning systems — where parts of the algorithm's behaviour are learned from data rather than written line by line — humans set the goals the system optimises for, select the training data, and decide how the system is deployed (Metzler & Garcia, 2024; Narayanan, 2023). The outputs of an algorithm therefore reflect the choices, assumptions, and values of those who build and deploy it; they are not neutral or objective.\n\nIn digital platforms, algorithms are typically combined into **algorithmic systems** — applications that, often using mathematical models, gather, combine, sort, classify, and infer data, and that select, prioritise, recommend, or make decisions on that basis (Council of Europe, 2020).\n\n\n:::info\n* UNESCO (2017). A lexicon for the digital age. UNESCO Courier, July–September 2017. \n* Committee of Ministers of the Council of Europe (2020). Recommendation CM/Rec(2020)1 on the human rights impacts of algorithmic systems, adopted 8 April 2020.\n* Metzler, H., & Garcia, D. (2024). Social drivers and algorithmic mechanisms on digital media. Perspectives on Psychological Science. \n* Narayanan, A. (2023). Understanding social media recommendation algorithms. Knight First Amendment Institute, Columbia University. \n\n:::","HTML":"

Algoritmi

\n

Algoritam je niz uputa koje računalo automatski izvršava (UNESCO, 2017.). On pretvara ulazne podatke u izlazne izvršavajući izračune, sortirajući ili klasificirajući podatke, donoseći odluke ili automatizirajući ponavljajuće zadatke. Algoritmi se kreću od vrlo jednostavnih (npr. fiksno pravilo za sortiranje popisa) do vrlo složenih (npr. sustav preporuka na društvenoj mreži).

\n

Algoritme osmišljavaju ljudi (Vijeće Europe, 2020.). Čak i u modernim sustavima strojnog učenja — gdje se dijelovi ponašanja algoritma uče iz podataka, a ne pišu se redak po redak — ljudi postavljaju ciljeve koje sustav optimizira, odabiru podatke za obuku i odlučuju o načinu implementacije sustava (Metzler & Garcia, 2024; Narayanan, 2023). Isputi algoritma stoga odražavaju izbore, pretpostavke i vrijednosti onih koji ga grade i primjenjuju; oni nisu neutralni ni objektivni.

\n

Na digitalnim platformama algoritmi se obično kombiniraju u algoritamske sustave — aplikacije koje, često koristeći matematičke modele, prikupljaju, kombiniraju, sortiraju, klasificiraju i izvode zaključke iz podataka te na temelju toga odabiru, postavljaju prioritete, preporučuju ili donose odluke (Vijeće Europe, 2020.).

\n
\n
\n\n
\n
\n
","UPDATEDAT":"2026-05-08T12:38:41.584Z","LANG":"hr","ID":"79384650-1b43-4959-b041-9f7ee6421ff5","TITLE":"Rječnik","SOURCELANG":"en"}