AI teaching assistants

 As we embark on the next phase of our AI project, building Gen AI agents to support our learners, a look at what is out there is important. This site by Educators Technology (published June 2025) has a lists of top AI teaching assistants. 

The site provides a definition of AI teaching assistant as a tool to help with lesson planning, grading, content creation, student feedback, and classroom management. Routine work including real-time support to learners, instruction through automation, personalisation of content and conversational interactions with learners can also be supported.

24 teaching assistants, some of which are able to do many of the items listed above, and some more specialised ones are listed. Their functions are summarised. Many require payment. Some tech companies are leveraging technology with the inclusion of AI into established models, others are new and many perhaps more suited to the school sector (and the U S of A) rather than higher or vocational education. Most have no links, so a search is required to find them. 

The discussion on AI teaching assistants needs to always consider their efficacy and whether they will replace human teachers. A June interview by Business Insider with Sal Khan, founder of Khan Academy; A radio NZ overview from late last year; and an August post from Education Week (US of A context), all indicate that teacher relationships and presence are crucial to good teaching. What AI TAs can do, is provide 24/7 support on FAQs, along with providing a place for students to undertake practice or continued learning. 

Therefore, using AI TA's needs to be carefully thought through. They should not replace teachers, but support teaching and learning. Identifying the components of the AI TAs role is a crucial step in making sure that learning objectives are met. Interconnecting the AI TAs role to that of the teacher, and making this visible to learners, is also important. Therefore, one aspect of AI literacy is to be able to understand how, when and why AI is used and what is the role of humans when AI takes on various roles.                                                                        

ChatGPT - study mode, Google guided learning, Claude learning mode and University of Sydney Cogniti - are they similar or different?

 There has been a flurry of activity in the Gen Au space of relevance to teaching and learning. First up was the launch of ChatGPT5.0 which allows for the use of it in 'study mode', This allowsa for a shift in the emphasis of using Gen AI to 'provide answers' towards using it as a 'study coach'. 

A few days later, Google also joined the move with its 'guided learning' in Gemini. 

Whilst, Claude has provided a learning mode for some months.

The above join University of Sydney's Cogniti as possibilities for teachers and learners to move towards personalised learning environments. There is also a recent start up - Wild Zebra - which provisions personalised tutors to students. 

However, as with all the 'vanilla' Gen AIs, each has things it does well and things it will struggle with. 

For example, here is a comparison of ChatGPT's study mode with Claude learning mode by Toms Guide. 

Using chatgpt to compare university of sydney cogniti with chatgpt study mode yields some differences.

Key Differences Summarized:

Feature	Cogniti	ChatGPT Study Mode
Focus	Educational context, feedback, integration	General problem-solving and learning
Integration	Canvas and other learning platforms	General use
Accessibility	Equitably available to all students	Requires access to ChatGPT
Feedback	Personalized and standardized	Interactive and conversational
Tracking	Tracks student-AI interaction	Does not specifically track
Bias	Potential for bias from training data	Potential for bias from training data

Therefore, each tool has pluses and minuses and as per all of our recent studies into  integration of Gen AI into VET, Gen AI tools need to be carefully selected, and learning planned and structured. 

A caveat with using Gen AI systems as 'tutors' is provided in a recent article by Flenady and Sparrow (2025). Their warning points to the often disregarded conceptualisation of Gen AI - in that it is NOT intelligent but build on algorithms for pattern recognition. They argue that Gen AI systems are 'epistemically irresponsible'. It is therefore important to always take heed of this warning and to ensure that all users have this at the topmost of their minds whenever they use Gen AI.

Gen AI and education: Digital pedagogies, teaching innovation and learning design - springer brief - book overview

This book  by Professor B. Mairēad Pratschke, arrived late last week, just in time for a wet weekend. It is a Springer brief of just over 100 pages, making it a quick but worthwhile read.

The book has 7 chapters. 

Chapter 1 covers the historical evolution of AI in education, a good overview of Gen AI (GAI is used in the book) and a summary of the tenets of digital pedagogy (including connectivism, social constructivism). The chapter closes with the call and rationale for all educators to attain GAI competencies and that its introduction into learning must be teacher-led, not as directives from above.

The second chapter undertakes the unpacking of the implications of GAI on the learning ecosystem. The term AI ecosystem is used to describe how GAI is integrated into teaching platforms and tools, the implications for undertaking this type of innovation, and provides some indications of how AI will impact on education as a whole. We are already able to customise AI bots but in the very near future, embodied AI and the opportunity to not just integrate GAI but to create new forms of education.

In the third chapter, the concept of human-computer hybrid is explored. In doing so, humans (and AI) draw on each other's strengths, leading to true synergy between biological/human and digital/machine/AI. The developments already available are summarised - including ways Gen AI is trained, the role of prompt engineering, Khan academy's experiments with personalised learning using AI, Poe and Hugging Face as the options to create your own bot, expert systems and examples in deploying Retrieval-Augmented Generation (RAG). A refresh of TPACK is proposed to become TPAIK, whereby 'knowledge' is shifted to be 'intelligence'. This leads to thinking about the relationships between technology/ntelligence/knowledge and how these are part of the learning design. Who creates intelligence/knowledge and what is the role of technology in the co-relationship?

In chapter 4, learning design to bring about the connections between educational theory, and digital pedagogy and practice is proposed. The concept of Generativism is proposed to bring GAI into the design, delivery and assessment of learning. The ABC learning design framework (with the learning activities of acquisition, collaboration, discussion, investigation, practice and production from Laurrilard's work) with GAI added, is provided as an example of how GAI contributes to each of the learning activities.

Chapter 5 undertakes an exploration and discussion of the impact of GAI specifically on personalised and peer learning. Intelligent tutoring systems, adaptive learning platforms, integrated assistants and tutors, standalone assistants and tutor, along with aspects of social learning, the implications of affective computing, social AI, intelligent communities and collaborative learning (see Sharple's work) are all introduced, discussed and evaluated. Each has a role, the importance is in selection and emphasis and of the roles of AI supporter and learner. Who has autonomy and what is the role of the educator when personalised learning environments become the norm.

The next chapter focuses on assessing learning. Congruent to the findings from our own AI projects, the emphasis in the AI age, is to place importance on the process of learning, rather than the outputs of learning. The ABC framework is again used, to provide guidance as to how assessments change due to AI and how AI can be used to support the assessment process.

In the last chapter, 'embedding AI', the many themes and threads across the book are brought together. Actions for educators and administrators are provided to help move education into the GIA age. 

In sum, a good book providing a summary of what has occurred thus far with good advice on the way forward. The book is short and therefore not too daunting for the practitioner/teacher/academic leader/head of school etc. to read through. There are good discussion anchors throughout the book and these will be useful as the difficult conversations with on GAI's role in education begin to take place in earnest. 

Personalise learning in the trades - AI coach in workshop learning

The work of Asplund, Kilbrink and team from Karlstad University in Sweden has been contributing much needed micro analysis of how people learn a trade. This 2022 paper, teaching and learning how to handle tools and machines in vocational education workshop sessions, follows on from another 2022 paper on 'introducing the object of learning in interaction:vocational teaching and learning in a plumbing workshop session'. 

Both delve into the differences between VET and other tertiary studies, whereby the focus in VET is on tools/machines and learning often requires not only 'hands-on' repetitive deliberate practice, but also one on one (preferably) learning conversations between learner and teacher/mentor/workplace trainer.

As such, there is potential for the use of AI in personalised learning in situ. Whereas most of the work on personalised learning involves PC or mobile phones. Learners may undertake 'programmed learning' whereby the learner is taken through practice-based learning through text, multimodal or simulated learning or revise their learning through various personalised learning quizzes.

In VET learning undertaken in authentic workshops whereby learners are completing full-time 'pre-trade' programmes or apprentices are learning new skills during block courses. There will usually be a ratio of 16-20 learners to one tutor/trainer. Individualised sessions are often 'hit and miss' and many learners have to rely on their own learning instincts, to carry them through practice sessions.

The advent of AI provides for new opportunities. ChatGPT4o allows for context-aware f2f AI interaction and runs on mobile devices. When we tested it with ESOL learners, the AI picked up on the accent of the learners and greeted the learner in their native language. Therefore, with appropriate further training, the AI should be able to provide individualised coaching or 'study buddy' support to learners, who are not tethered to a PC. 

There are also wearable AI options as exampled in this Guardian article which reports on the AI Pin. A quick search reveal several similar options, including an AI pendant which 'keeps you company' 

Using such a device, would provide just in time feedback and support, when a learner needs assistance. Otherwise, and very often, the teachable/learning moment passes and the opportunity is lost. Something to consider for a future project :) 

Leaders and Legends Online learning - Professor Mike Sharples

 Dr. Mark Nichols' podcast on Leaders and Legends of Online Learning this week is with Professor Emeritus Mike Sharples. 

Began with an overview of his career, how it started with a degree in computer science and worked towards his PhD on cognition, computers and creative writing. He has always worked with AI and Gen AI for over 50 years. 

He compared the 'sudden awareness' of AI to how the internet developed. Much work undertaken over many decades before user relevance and other parts supporting the concept, took it into the mainstream. As with the www, important to think through implications and have guardrails with AI as there are many ways it can be utilised for good and also for bad.

Discussed his involvement across the 2000s with mobile learning as a means to provide more personal and individualised learning. 

Then provided an overview of how the UK version of MOOC - futurelearn - was set up to bring in good practice and pedagogy based on understanding of neuroscience into the design of futurelearn. A successful learning platform, distinctively difference from other MOOCs which are based mainly on lectures. In contrast, futurelearn stresses social learning opportunities along with the usual online learning platform mechanics.

Then introduced his most recent books including Story Machines: How computers have become creative writers.

Discussed the importance of social constructivist learning. Personalised learning is one piece but not the only one. There is still a need for learners to interact with their peers, teachers etc. to springboard and synthesise, discuss and defend their stance, weight up and evaluate their conceptual understanding. AI should not be only a technological tool but needs to be led by pedagogy. Encouraged the need to ensure the use of technology is more human centred. 

Summarised the important components of a new online system. Pedagogy is essential. Concepts include the need to have spaced learning, ensure learning is a social process, feedback is provided at the right time,and the need to build learner efficacy. AI-enhanced collaborative learning must be the goal.

Proposed the most important research is to find out how to best leverage technology to encourage and support social learning, rather than just go down the personalised learning route. New methods of assessment also need to be considered to allow for social learning and to focus on assessments for learning. Some ideas include peer assessments, the opportunity for learners to express their judgment, and evaluative and critical thinking. 

Now he is retired, his research interest centres around the future of technology-enhanced education. He is able to concentrate on research without the distractions of the other aspects of an academic career. Recommended to follow the work of colleagues at the University of Sussex and Looi Chee Kit (Nanyang Technological University and The Educational University of Hong Kong) who has had success introducing many principles of learning sciences at the primary school level.