AI systems are non-neutral socio-technical artifacts embedding cultural assumptions and power relations¶

Claim · All claims
Evidence · 1 study · theoretical r2 · q1
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q1 i? AI systems embed cultural assumptions, economic and political interests, and power relations, mediating human activity in distinctive ways. → Juliana Elisa Raffaghelli 2026

Evidence¶

Juliana Elisa Raffaghelli 2026¶

Juliana Elisa Raffaghelli, Henriikka Vartiainen, Matt Bower, Martina Ronci, Chris Shelton, Kathryn MacCallum, Jihyun Lee, Mary Webb, Yousra Cthouki, David Smith. (2026). Situated AI Ethics: A Cultural-Historical and Ecological Framework for Education. Computers and Education Open. https://doi.org/10.1016/j.caeo.2026.100368

q1 · i? · theoretical · r2

Theoretical assertion in the discussion, attributed by the article to Coeckelbergh, Crawford and Hopwood. The article illustrates that when under-resourced schools adopt ChatGPT as a standard tool, 'epistemic injustice may arise through the marginalisation of minority languages and perspectives'.

"AI systems are never neutral tools (Coeckelberg, 2023); they embed cultural assumptions, economic and political interests, and power relations (Crawford, 2021), mediating human activity in distinctive ways (Hopwood, 2023)."

Studies bearing on this claim¶

Do the assumptions built into an AI tool carry into what learners write and the feedback they get, and does configuring the tool change that? This table gathers the studies behind this claim and the 9 narrower claims that bear on it: 6 distinct studies (2 causal, 2 qualitative, 1 theoretical, 1 review), q1–q2, 1 of 6 with an effect size coded. One row per study, codes as each entry records them; no effect is averaged and no verdict is drawn. Built by scripts/synthesize_claim.py from eval/synthesis/; a row that looks wrong is fixed on the claim it comes from.

Discussion¶

Related Claims¶

• Existing institutional AI literacy frameworks converge on a consumer model with no substantive treatment of training data politics or governance participation — related
• Community projects need conceptual framing to avoid narrowing the learning object — related
• Conceptions of how children learn literacy are cultural and political as much as scientific, shaping different classroom practices across countries — related
• The agent/consumer distinction is not primarily technical sophistication but capacity to interrogate the assumptions of AI outputs — related
• Storying is socially and politically non-neutral: it privileges some storylines and silences others — related
• Deepseek's justification scores vary by student socio-economic profile, with Top Tier students averaging 9.60 versus 6.90 for Low Tier students — a narrower finding that bears on this claim
• Algorithmic empathy is culturally fragile: emotional inference errors in socio-emotional AI can invalidate teachers' experiences, so systems require local validation and participatory design — a narrower finding that bears on this claim
• Cultural mediation has four implications, including that artifacts fundamentally transform mental functioning and the mind is distributed in artifacts — a broader claim this one bears on
• Human-AI collaboration requires refined group-awareness concepts for asymmetrical relationships between users and AI partners — a narrower finding that bears on this claim
• Primary barriers to infusing 21st century skills are psychological, political, and cultural rather than conceptual, technical, or economic — a narrower finding that bears on this claim
• No reviewed participatory design method satisfies all seven criteria for designing disruptive socio-technical systems — a narrower finding that bears on this claim
• Despite a student's claims of purely human authorial agency, posthumanist analysis of his AI-mediated revision shows meaning co-produced through human-AI entanglement — related
• Unstructured LLM use contributes to epistemic narrowing by privileging convergent outputs and dominant viewpoints — a narrower finding that bears on this claim
• A gender-biased system prompt induces large gender-differentiated language in LLM-generated career essays, while a neutral prompt does not — a narrower finding that bears on this claim
• Shared technology resources can act as positioning artifacts around which students negotiate power and status differentials in CSCL groups — a narrower finding that bears on this claim
• Underlying assumptions dictate how the five foundations are operationalized in a learning environment — a narrower finding that bears on this claim
• All AI-generated outputs in the project underwent mandatory human-in-the-loop review, positioning AI as augmentative rather than substitutive — a narrower finding that bears on this claim
• Educators cautiously support AI for culturally responsive content generation while fearing cultural biases and stereotypes in AI output — related
• Students assisted by a gender-biased LLM produce career-plan essays with a significantly larger male-female agentic gap than students in neutral-LLM and no-AI control conditions — a narrower finding that bears on this claim
• Assessment instruments are never neutral but always mediational, embedding a culturally biased telos — a narrower finding that bears on this claim