Experimental-group students detected AI errors and enacted verification and ethical self-regulation as routine practices¶

Claim · All claims
Evidence · 1 study (2 entries) · causal r2 · q1
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q1 i? Students recognized hallucinated or fabricated AI content and verified AI claims through systematic triangulation with external sources, with verification becoming a habitual practice (Theme 2 in 100% of reflections; Theme 3 in 81%). → Alshehri OAO 2026 q1 i? Students articulated concrete self-regulation strategies for future AI use, including drafting independently first, restricting AI functions, and setting time limits (Theme 4, 100% of reflections). → Alshehri OAO 2026 (2)

Evidence¶

Alshehri OAO 2026¶

Alshehri OAO, Lamouchi A, Abdellatif MS, Al-Dosari MNA, Mekheimer M and Nemt-allah MA (2026). From tool to scaffold: structured human–AI collaboration and its effects on academic writing and digital critical thinking among Saudi EFL learners. Front. Psychol. 17:1830103. https://doi.org/10.3389/fpsyg.2026.1830103

q1 · i? · causal · r2

Qualitative thematic analysis of experimental-group reflections (Theme 3, verification as digital critical practice, evidenced in 81% of reflections per Table 5). Theme 2 additionally documented hallucinated citations and cultural or tonal mismatches students identified and repaired.

"Students described cross-checking AI claims via Google Scholar, library databases, reputable news outlets, and official websites, retaining only content that could be externally supported."

Alshehri OAO 2026 (2)¶

Alshehri OAO, Lamouchi A, Abdellatif MS, Al-Dosari MNA, Mekheimer M and Nemt-allah MA (2026). From tool to scaffold: structured human–AI collaboration and its effects on academic writing and digital critical thinking among Saudi EFL learners. Front. Psychol. 17:1830103. https://doi.org/10.3389/fpsyg.2026.1830103

q1 · i? · qualitative · r?

Qualitative thematic analysis of the same reflections (Theme 4, metacognitive regulation and ethical self-positioning, 100% of reflections). Students positioned AI as a tool, emphasized authorship maintenance, plagiarism avoidance, and alignment with course expectations.

"Looking forward, students outlined specific self-regulation strategies: drafting independently before using AI, restricting AI to defined functions such as grammar checks and vocabulary refinement, setting time limits on AI interaction, and verifying sources immediately rather than near submission."

Discussion¶

Related Claims¶

• The most frequently cited ethical concerns about GenAI in research were hallucinated facts, fabricated references, and misleading content — related
• Students' top two concerns about educational LLM use are output quality (fabricated or misleading content, 53 mentions) and the verification burden (34 mentions) — related
• Students report negative experiences with ChatGPT including inaccurate outputs, lack of emotional connection, and risk of overreliance, which constrain competence — related
• Students remained cautious about relying on GenAI throughout the creative process, citing imperfect outputs and perceived loss of ownership — related
• Student experiences with GenAI in design thinking fall into four themes: perceived benefits, ethical concerns, hesitance and acceptance, and critical validation — a broader claim this one bears on
• Source provenance labels functioned as a seal of authority: 11 of 12 participants failed to detect a planted source-card mismatch — related
• GenAI improves administrative coordination of group work but its analyses of team members can be inaccurate and hallucinated — related
• Pilot students reported structural limitations of AI-SEE: AI-tool entry barriers, limited hands-on time, and partial disconnection between ethics instruction and routine engineering decisions — reports the opposite
• AI-generated ethics exercise materials carried external citations that did not survive checking, ranging from misattributed to apparently fabricated — related
• Experimental-group students framed AI as a flexible writing scaffold supporting their own composing rather than a substitute for authorship — related
• Checking can coexist with unresolved errors: 7 of 15 students failed a task containing plausible LLM falsehoods despite self-checking — reports the opposite
• Interviews revealed six themes: enhanced efficiency, improved quality, facilitated reflection, increased interaction, improved collaboration, and concerns about over-reliance — related
• Surface validation can function while substantive validation fails: most participants detected flawed AI recommendations yet continued following them — reports the opposite
• Review-attributed evidence: fabricated model-generated references clustered in disciplinary journals in a student verification exercise — related
• Adolescent digital self-regulation is enacted through a typology of four strategy types differing in executive effort versus external scaffolding — related