A critical and inventive orientation to algorithms that foregrounds their alterity—their capacity to produce otherwise—by attending to how data, models, and interfaces configure relations, and by designing alternative algorithmic practices that redistribute agency, accountability, and care.
Altergorithm names a way of engaging algorithms that treats them as mutable, situated, and ethically charged socio-technical practices rather than neutral procedures. Drawing on posthumanist and new materialist commitments in Braidotti, Haraway, and Barad, it invites attention to the apparatus-level decisions that enact what data becomes, how training corpora sediment histories, which features are amplified, and whose futures are anticipated. In this register, algorithms are not simply tools that process inputs into outputs; they are material-discursive configurations that cut worlds together-apart, entangling humans, machines, infrastructures, and institutions in patterned ways. Altergorithmic work asks how those patterns might be reconfigured so different relations, accountabilities, and possibilities can take root.
The “alter” in altergorithm signals more than critique. It orients toward invention, toward composing alternative algorithmic practices that surface lineages, uncertainties, and exclusions, and that cultivate response-ability in use. It sits alongside critical data studies and platform studies while drawing energy from feminist technoscience’s ethos of “staying with the trouble” in Haraway and the affirmative ethics of Braidotti. With Barad’s agential realism, altergorithm emphasizes that model architectures, prompting conventions, and interface affordances are part of the phenomenon; they enact agential cuts with ethical consequence. Altergorithmic practice therefore experiments with ways of prompting, fine-tuning, documenting, and governing that make these cuts discussable, contestable, and revisable.
Postqualitative inquiry treats methods as worlding practices. Altergorithm fits this stance by treating algorithmic components as elements of the research apparatus whose configurations shape what is sensed, recorded, and analyzed. It foregrounds pipeline transparency (data provenance, pre-processing, prompting schemas), traceable decision points (hyperparameters, sampling, filtering), and situated validation (fit-to-purpose judged with those affected). Analytically, it supports diffractive readings across code, logs, outputs, and field materials to follow how algorithmic relations produce differences that matter. Ethically, it emphasizes response-ability: documenting exclusions and uncertainties, designing for consentful data use, and composing accounts that hold platform infrastructures and institutional policies in view, not only user behavior.
In AI-entangled writing instruction, altergorithm reframes classroom encounters with LLMs as opportunities to redesign relations among prompts, datasets, outputs, and assessment practices. It supports an instructional ethos that treats prompting as part of the composing apparatus whose configuration shapes authorship, voice, and originality. Altergorithmic assignments can make model lineage visible (training data disclosures, bias probing), render uncertainty explicit (calibration, rationales, version histories), and cultivate process accountability (prompt journals, decision rationales, citations of model contributions). This orientation invites the pedagogical assemblage to shift from detection and policing toward negotiated practices of attribution, consent, and co-composition, where responsibilities are distributed among students, instructors, platforms, and policies. It also aids in attuning to affective dynamics—relief, anxiety, curiosity—that arise when model outputs feel authoritative; altergorithmic practice designs interfaces and routines that slow down credulity, foreground limits, and keep space for human judgment and inquiry.