With the omnipresent and omniscient AI tools today, it is easier to identify education aspects that are not susceptible to be mixed with the new digital technologies. A variety of possibilities to account in education management and administration, many options for teachers professional development, a diversity of learning scenarios embedding AI tools, along with the pressure to change the syllabus to accommodate new ways of relating to knowledge and to learning – all these indicate the beginning of a shift in education that we will soon witness.

However, the use of AI didn’t start with ChatGPT moment, some education institutions previously exploring the use of advanced technologies for various purposes such as adaptive assessment, predictions, tutoring, educational games, and assistive systems. Today, there are important steps in many directions:

• learning – providing fast and precise answers to questions, helping students to improve their writing skills, providing immediate feedback, supporting test preparation, assisting experiments, investigations, topics exploration, argumentative thinking etc.
• adaptive learning/ personalisation of learning – AI can help tailor the content, pace and feedback to each student’s needs and preferences
• lesson planning/ learning paths design – finding relevant resources and activities, generate alternative lesson plans, providing feedback and suggestions
• tutoring – simulate one-on-one tutoring sessions with students and provide hints, explanations and scaffolding
• educational games – create engaging and immersive learning environments that can motivate and challenge students
• assistive technologies – increasing access, enhancing independence, boosting confidence
• early childhood education – adaptive games to train basic skills, engaging and effective gamified learning activities, personalised learning
• language learning – interactive practice, personalised feedback, adaptive learning
• plagiarism detection – adaptive text similarity analysis, paraphrasing detection, AI content detection
• exam integrity – text generation detection (identify interventions of AI writing tools), online proctoring, plagiarism detection
• assessment and grading – measure the student’s progress and skills in real-time and provide adaptive feedback and guidance
• automated feedback to students
• learning analytics – collect and analyse data from students’ interactions with learning platforms and provide insights for improving learning outcomes
• evaluation of the course (and QA) – provide automated reports based on the results of satisfaction evaluation (immediate feedback from students), knowledge evaluation, impact evaluation
• classroom/ behaviour management – attendance, attention monitoring, behaviour tracking, personalised learning, communication with parents
• scheduling – timetable, transportation, maintenance & administration of supplies
• predictions - e.g. students at risk of falling behind, dropout, etc.
• educational institutions management and administration – decision support based on data analysis and machine learning, administrative efficiency, professional development
• safety/ cybersecurity – threat detection and threat prevention, rapid complex response to threats

In particular in higher education, there are areas in which advanced universities made a custom of using assistive AI technologies to support essential activities and make them more effective:

• chatbots for enrolment and retention
• enhanced online discussion boards
• complex learning management systems
• connecting campuses
• analysing student success metrics
• transcription of faculty lectures
• inclusion and universal access (translations, automatic text reading, closed captions, remote access to synchronous activities, matching credits from other learning paths etc.)
• individualised learning
• plagiarism detection
• exam integrity
• academic research.

An interesting and pragmatic approach is proposed by UNESCO (2023) to support the decision of choosing technology in education: From the radio right through to generative AI, when and how should we use technology in education?When does it support learning and when does it distract? Do we need to change what and how we are learning with technology to keep education relevant? Four questions are to be answered before employing any new tool in our teaching and learning practices:

(1) Is it appropriate? ”No single device will improve learning everywhere for everyone. Some technology can improve some types of learning in some contexts. But the appropriate technology must be chosen for your particular case. […]”

(2) Is it equitable? ”[…] Our focus must remain on the most marginalized as we invest.”

(3) Is it scalable? ”[…] Deciding on whether to scale-up our choices means having a grasp of their long-term costs, but many ignore or are unaware of the full picture.”

(4) Is it sustainable? ”[…] (We) must map out and prioritize the digital skills they want to teach learners and teachers that will stand the test of time.”

In its Ethical Guidelines (2022), the European Commission presented a structured range of relevant use cases regarding the presence of AI in the teaching, learning, and assessment process:

1. STUDENT TEACHING. Using AI to teach students

Intelligent tutoring system	The learner follows a step-by-step sequence of tasks and gets individualised instruction or feedback without requiring intervention from the teacher.
Dialogue-based tutoring systems	The learner follows a step-by-step sequence of tasks through conversation in natural language. More advanced systems can automatically adapt to the level of engagement to keep the learner motivated and on task.
Language learning applications	AI-based learning apps are used in formal and non-formal education contexts. They support learning by providing access to language courses, dictionaries and provide real-time automated feedback on pronunciation, comprehension and fluency.

1. STUDENT SUPPORTING. Using AI to support student learning

Exploratory learning environments	Learners are offered multiple representations that help them identify their own routes to achieving the learning goals.
Formative writing assessment	Learners are provided with regular automatic feedback on their writing/ assignments.
AI-supported collaborative learning	Data on each learner’s work style and past performance is used to divide them into groups with the same ability levels or suitable mix of abilities and talents. AI systems provide inputs/suggestions on how a group is working together by monitoring the level of interaction between group members.

1. TEACHER SUPPORTING. Using AI to support the teacher

Summative writing assessment, essay scoring	AI is used to evaluate and grade learners’ written work automatically. AI and machine learning techniques identify features such as word usage, grammar and sentence structure to grade and provide feedback.
Student forum monitoring	Key words in student forum posts trigger automatic feedback. Discussion analytics provide insights to students’ forum activity and can highlight students who may need help or are not participating as expected.
AI teaching assistants	AI agents or chatbots provide answers to commonly asked questions by learners with simple instruction and directions. Over time, the AI system is able to broaden the range of answers and options provided.
Pedagogical resource recommendation	AI recommendation engines are used to recommend specific learning activities or resources based on each student’s preferences, progress and needs.

1. SYSTEM SUPPORTING. AI to support diagnostic or system-wide planning

Educational data mining for resource allocation	Universities/ schools gather student data which are analysed and used to plan how available resources can be best allocated for tasks like creating class groupings, assigning teachers, timetabling, and highlighting students who may require additional learning support.
Diagnosing learning difficulties	Using learning analytics, cognitive skills such as vocabulary, listening, spatial reasoning, problem-solving, and memory are measured and used to diagnose learning difficulties, including underlying issues that are hard for a teacher to pick up but might be detected early using AI systems.
Guidance services	AI based guidance services provide ongoing prompts or choice to create pathways for future education. Users can form a competence profile including previous education and include their own interests. From this data, combined with up-to-date course catalogue or study opportunity information, relevant study recommendations can be created using natural language processing.

Examples of frequently used AI tools for learning and researching

From many AI tools that can be explored in 2023, we have made a selection and we invite you to try them in order to get a good grasp of what students are doing or could be doing. Most of them are useful and should be recommended to students as learning companions, to improve their productivity, to motivate them to be smarter and more creative, to support their preparation for tomorrow’s professions:

• book search using natural language prompts: Talk to Books (Google)
• literature research, personalized alerts, etc.: ResearchRabbit (Human Intelligence Technologies Inc.)
• evaluation of scientific works – analysis of the topics covered by comparison with other relevant works: scite (Scite Inc.)
• finding relevant papers on a topic or starting from a research hypothesis: Elicit (Ough Inc.)
• pdf document analysis and synthesis; answer questions based on the material/ book: ChatPDF (Mathis Lichtenberger und Moritz Lage GbR)
• development of a scientific/ research article: Smodin (Smodin LLC) – paid. Instrument: Smodin Author (AI Writer).
• analysing a (course) video and creating the summary, reading notes, quiz/ flashcards based on the content: YouLearn (Achyut Byanjankar, Advait Paliwal, David Yu)

Planning for effective use of IA and data in educational institutions

The Ethical Guidelines prepared by the European Commission (2022) put basis for a proper implementation of AI, as part of the institutional digitalisation process and transformation towards more efficient administrative and teaching processes.

The first preparatory actions should be related to raising awareness and community engagement – as the most important aspect is preparedness of human resources: staff, collaborators, beneficiaries:

Discuss with colleagues. Collaboration between educators contributes to school/ university improvement and student success. Educators often draw support from each other and can delegate tasks in ways that help them collectively to be more effective. Working collaboratively can help to make more informed decisions and helps ensure a more consistent approach to using AI and data systems across the educational institution.

Collaborate with other educational institutions. Collaboration between schools/ universities is an effective way to share experiences and best practices and learn how other educational institutions have implemented AI systems. This can also be useful in identifying and dealing with reliable providers of AI and data systems that adhere to the key requirements for trustworthy AI. It is important that educational institutions participate in supervised projects and experiments organised at regional, national, or European level through initiatives such as Erasmus+. These provide opportunities for educators and decision makers to participate collaboratively in a process of applied research and inform future use and development of AI and data use in educational institutions.

Communicate with stakeholders. Involving all stakeholders and especially learners in discussions and decision making will lead to better understanding and trust in what the institution is aiming to achieve through the use of AI systems. Careful consideration needs to be given to explain what data is being collected, what is being done with the data, how and why it is being collected, and how this is protected. It will be important to share these explanations with learners (and parents) and to provide opportunities for them to provide their feedback and voice possible concerns. Learners might require different approaches in order to engage them so that they can participate in informed decision making.

Keep up to date. As AI systems continue to evolve and data usage increases, it is very important to develop a better understanding of their impact on the world around us, including in education and training. Educators will need to continue to keep informed of new innovations and development through participation in continuing professional learning and involvement in communities of practice. Decision makers will need to provide opportunities for staff to upskill and continue to develop competences for ethical use of AI and data. – (European Commission, 2022)

The same Ethical Guidelines suggest the main steps to be approached when planning for effective use of AI in educational institutions (EC, 2022):

Review current AI systems and data use	Research about the AI systems that are already in place. When carrying out a review, it is useful to list what data is being gathered by the educational institution and clarify what purpose this serves. Consider how long the data will be needed for and how the school/ university might be able to retain it for as little time as possible. The European Union General Data Protection Regulation (GDPR) requires this kind of analysis.
Initiate policies and procedures	Prior to implementing an AI system, institutional wide policies and procedures need to be put in place to establish expectations and to provide guidance on how to consistently deal with issues when they arise. These could include measures for: • ensuring public procurement of trustworthy and human-centric AI; • implementing human oversight; • ensuring that input data is relevant to the intended purpose of the AI system; • the provision of appropriate staff training; • monitoring the operation of the AI system and taking corrective actions; and, • complying with relevant GDPR obligations, including carrying out a data protection impact assessment.
Carry out a pilot of the AI system	It can be useful to trial the system with a particular learner cohort. It is important to have a clear vision of what is to be achieved with the new technology so that an informed decision can be made involving students and other stakeholders. Specific evaluation criteria are required so that an informed judgement can be made on the effectiveness of the AI system in terms of improvement of learning outcomes, value for money and ethical use. This will also highlight some of the key questions that may need to be asked of the supplier before purchasing the system.
Collaborate with the AI system provider	It is important to maintain contact with the AI system provider prior to deployment and throughout the lifecycle of the AI system. Look for clear technical documentation and seek clarification on any aspects that are unclear. Assurances should be sought from the provider in terms of their adherence to applicable legal obligations. The university/ school should also consider future dependence on the provider if, for example, it seeks to change provider in the future, or move to a different AI system altogether. It is also important that any human oversight measures identified by the provider are implemented by the educational institution while the AI system is being used.
Monitor the operation of the AI system and evaluate the risk	The use of the AI system should be monitored on an ongoing basis to evaluate the impact on learning, teaching, and assessment practices. At school/ university level, it will be important to decide how monitoring will be organised and carried out, who will be responsible for monitoring and how progress will be determined and reported. The evidence gathered, as a result of ongoing monitoring, should inform and influence the future use of AI systems or the decision not to use them in particular circumstances.