User:Kevinmount/Educational technology

Edited Artificial Intelligence Section
The history of artificial intelligence initiated in 1956 by cognitive scientists, focused on investigating thought and learning processes in humans and machines. The earliest uses of AI in education can be traced back to the development of intelligent tutoring systems (ITS) and their application in enhancing educational experiences. They are '''designed to provide immediate and personalized feedback to students. ''' The incentive to develop ITS comes from educational studies showing that individual tutoring is much more effective than group teaching, in addition to the need for promoting learning on a larger scale. Over the years, a combination of cognitive science theories and data-driven techniques have greatly enhanced the capabilities of ITS, allowing it to model a wide range of students' characteristics, such as knowledge, affect, off-task behavior and wheel spinning. There is ample evidence that ITS are highly effective in helping students learn. ITS can be used to keep students in the zone of proximal development (ZPD): the space wherein students may learn with guidance. Such systems can guide students through tasks slightly above their ability level.

'''Generative Artificial Intelligence (GenAI) emerged with the introduction of ChatGPT in November 2022.  This caused alarm among K-12 and higher education institutions, with a few large school districts quickly banning GenAI, due to concerns about potential academic misconduct. However, as the debate developed, these bans were largely reversed within a few months. To combat academic misconduct, detection tools have been developed, but are neither accurate nor reliable. '''

There have been various use cases in education, including providing personalized feedback, brainstorming classroom activities, support for students with special needs, streamlining administrative tasks, and simplifying assessment processes. However, there are concerns that GenAI's can output incorrect information, also known as hallucination. The results from GenAI can also be biased, leading to calls for transparency regarding the data used to train GenAI models and their use. Providing professional development for teachers and developing policies and regulations can help to mitigate the ethical concerns of GenAI. And while AI systems can provide individualized instruction and adaptive feedback to students, they have the potential to impact students' well-being and sense of classroom community.

Copied from Artificial Intelligence (original) - Small Edit changes bolded
As artificial intelligence (AI) becomes more prominent in this age of big data, it has also been widely adopted in K-12 classrooms. One prominent class of AI-enhanced educational technology is intelligent tutoring systems (ITSs), designed to provide immediate and personalized feedback to students. The incentive to develop ITS comes from educational studies showing that individual tutoring is much more effective than group teaching, in addition to the need for promoting learning on a larger scale. Over the years, a combination of cognitive science theories and data-driven techniques have greatly enhanced the capabilities of ITS, allowing it to model a wide range of students' characteristics, such as knowledge, affect, off-task behavior and wheel spinning. There is ample evidence that ITSs are highly effective in helping students learn. ITSs can be used to keep students in the zone of proximal development (ZPD): the space wherein students may learn with guidance. Such systems can guide students through tasks slightly above their ability level.

Recent works have also focused on developing AI-enhanced learning tools that support human teachers in coordinating classroom activities. The advantages of AI education include personalized learning, support for students with special needs, streamlined administrative tasks, and simplified assessment processes. The teacher can support students in a way that AI cannot, but is unable to process the large amount of real-time data analytics provided by the computer system. On the other hand, AI can share the workload and recommend the best course of action (e.g., by pointing out which students require the most help), but can only operate in the pre-specified domain and cannot handle tasks such as providing emotional support or remedial lessons to students in need.  However, existing systems were designed under the assumption that students progress at the same pace . Understanding how to support teachers in a realistic, highly differentiated, self-paced classroom, remains an open research problem.