AI-Powered Educational Agents: Opportunities, Innovations, and Ethical Challenges DOI Creative Commons
Diana‐Margarita Córdova‐Esparza

Information, Год журнала: 2025, Номер 16(6), С. 469 - 469

Опубликована: Май 31, 2025

Recent advances in large language models (LLMs) have triggered rapid growth AI-powered educational agents, yet researchers and practitioners still lack a consolidated view of how these systems are engineered validated. To address this gap, we conducted systematic literature review 82 peer-reviewed industry studies published from January 2023 to February 2025. Using four-phase protocol, extracted coded them along six groups: technical pedagogical frameworks, tutoring systems, assessment feedback, curriculum design, personalization, ethical considerations. Synthesizing findings, propose design principles that link choices instructional goals outline safeguards for privacy, fairness, academic integrity. Across all domains, the evidence converges on key insight: hybrid human–AI workflows, which teachers curate moderate LLM output, outperform fully autonomous tutors by combining scalable automation with expertise. Limitations current literature, including short study horizons, small-sample experiments, bias toward positive temper generalizability reported gains, highlighting need rigorous, long-term evaluations.

Язык: Английский

AI-Powered Educational Agents: Opportunities, Innovations, and Ethical Challenges DOI Creative Commons
Diana‐Margarita Córdova‐Esparza

Information, Год журнала: 2025, Номер 16(6), С. 469 - 469

Опубликована: Май 31, 2025

Recent advances in large language models (LLMs) have triggered rapid growth AI-powered educational agents, yet researchers and practitioners still lack a consolidated view of how these systems are engineered validated. To address this gap, we conducted systematic literature review 82 peer-reviewed industry studies published from January 2023 to February 2025. Using four-phase protocol, extracted coded them along six groups: technical pedagogical frameworks, tutoring systems, assessment feedback, curriculum design, personalization, ethical considerations. Synthesizing findings, propose design principles that link choices instructional goals outline safeguards for privacy, fairness, academic integrity. Across all domains, the evidence converges on key insight: hybrid human–AI workflows, which teachers curate moderate LLM output, outperform fully autonomous tutors by combining scalable automation with expertise. Limitations current literature, including short study horizons, small-sample experiments, bias toward positive temper generalizability reported gains, highlighting need rigorous, long-term evaluations.

Язык: Английский

Процитировано

0