Table 1 Evolution of AI Phase Period Key... | Emerald Publishing

Table 1

Evolution of AI

Phase	Period	Key AI technologies/concepts	Impact on education and pedagogy	Associated challenges and educator perceptions
Phase 1: Foundational AI and early educational tools	1950s – 1980s	Alan Turing’s foundational concepts, ELIZA (early NLP), computer-assisted instruction (CAI) (e.g. PLATO and TICCIT), rule-based intelligent tutoring systems (ITS) (e.g. cognitive tutor) and the influence of Skinner and Bloom	Introduction of automated, individualised instruction; self-paced learning; immediate feedback; focus on mastery learning; positive student attitudes and learning rates, especially for disadvantaged students	High costs and limited access to bulky technology; system inflexibility (rule-based, limited adaptability); mixed effectiveness for deeper learning; ITS lacked human nuance/emotional touch; initial educator reluctance and unpreparedness; historical resistance to new tech (e.g. calculators)
Phase 2: Machine learning and adaptive systems	1990s – 2010s	Advancements in machine learning (ML) and natural language processing (NLP); “adaptive AI” (e.g. knowledge tracing); integration into learning management systems (LMS); rise of massive open online courses (MOOCs) (e.g. Coursera and edX); AI for data analysis and research support	Enhanced personalisation and adaptive learning at scale; faster identification of learning gaps; data-driven insights for curriculum optimisation; increased accessibility of quality education; and streamlined administrative tasks	Digital divide persistence; interoperability issues and need for high-quality data; educator digital literacy gaps and resistance (e.g. internet and Wikipedia); “adaptive loops of death” in early systems; ITS still lacked social/emotional engagement; and general feeling of unpreparedness among educators; emerging data privacy concerns
Phase 3: Deep learning and GenAI revolution	2010s – Present	Deep neural networks; large language models (LLMs) (e.g. GPT-2, GPT-4 and ChatGPT); GenAI for multimodal content (e.g. DALL·E and Midjourney); and AI-powered assistants and autonomous agents	Unprecedented levels of personalisation and adaptive feedback; significant automation of teacher content creation and administrative tasks; new opportunities for inquiry-based learning and creative exploration; data-driven curriculum evolution; and enhanced accessibility and equity potential	Paramount academic integrity and plagiarism concerns (flawed AI detection); heightened data privacy and security risks; risk of student over-reliance and critical thinking erosion; widespread educator unpreparedness and lack of institutional guidance; algorithmic bias and ethical dilemmas; overwhelming product landscape and cost barriers; and resistance to fundamental pedagogical shifts

Phase	Period	Key AI technologies/concepts	Impact on education and pedagogy	Associated challenges and educator perceptions
Phase 1: Foundational AI and early educational tools	1950s – 1980s	Alan Turing’s foundational concepts, ELIZA (early NLP), computer-assisted instruction (CAI) (e.g. PLATO and TICCIT), rule-based intelligent tutoring systems (ITS) (e.g. cognitive tutor) and the influence of Skinner and Bloom	Introduction of automated, individualised instruction; self-paced learning; immediate feedback; focus on mastery learning; positive student attitudes and learning rates, especially for disadvantaged students	High costs and limited access to bulky technology; system inflexibility (rule-based, limited adaptability); mixed effectiveness for deeper learning; ITS lacked human nuance/emotional touch; initial educator reluctance and unpreparedness; historical resistance to new tech (e.g. calculators)
Phase 2: Machine learning and adaptive systems	1990s – 2010s	Advancements in machine learning (ML) and natural language processing (NLP); “adaptive AI” (e.g. knowledge tracing); integration into learning management systems (LMS); rise of massive open online courses (MOOCs) (e.g. Coursera and edX); AI for data analysis and research support	Enhanced personalisation and adaptive learning at scale; faster identification of learning gaps; data-driven insights for curriculum optimisation; increased accessibility of quality education; and streamlined administrative tasks	Digital divide persistence; interoperability issues and need for high-quality data; educator digital literacy gaps and resistance (e.g. internet and Wikipedia); “adaptive loops of death” in early systems; ITS still lacked social/emotional engagement; and general feeling of unpreparedness among educators; emerging data privacy concerns
Phase 3: Deep learning and GenAI revolution	2010s – Present	Deep neural networks; large language models (LLMs) (e.g. GPT-2, GPT-4 and ChatGPT); GenAI for multimodal content (e.g. DALL·E and Midjourney); and AI-powered assistants and autonomous agents	Unprecedented levels of personalisation and adaptive feedback; significant automation of teacher content creation and administrative tasks; new opportunities for inquiry-based learning and creative exploration; data-driven curriculum evolution; and enhanced accessibility and equity potential	Paramount academic integrity and plagiarism concerns (flawed AI detection); heightened data privacy and security risks; risk of student over-reliance and critical thinking erosion; widespread educator unpreparedness and lack of institutional guidance; algorithmic bias and ethical dilemmas; overwhelming product landscape and cost barriers; and resistance to fundamental pedagogical shifts