Artificial intelligence (AI), as powered and enhanced by big data enabling technologies such as the Internet of Things (IoT), a technology commercialised only within the decade leading up to 2025, is a popular technology, embedded in many products that are used in everyday services and applications.
AI technology is integrated with educational curricula across Europe, such as the robot that supports learning for younger students, known as lifelong learning companion. Whilst the term lifelong is currently used loosely, the learning companion joins the young student to support his/her learning experience only during elementary school. In addition, AI is also used for the transformation of classrooms into game-like environments to enhance student experience in schools, and to enhance their skillset.
AI is already a part of everyday life, often without consumers realising that they are dealing with AI-powered software products, not necessarily in the form of humanoid robots, but as AI-powered software integrated in several everyday products. For example, AI-powered bots have replaced service assistants in more than 80 per cent of customer service departments in the retail industry; these bots are using machine learning to respond to customer requests. The availability of huge amounts of data from many different sources has enabled educational technologies based on AI to flourish in 2025.
Educational technologies are redefining the role of the teacher to become a facilitator of the learning activity. This facilitation becomes needed, while simultaneously, demand for the content and information aspects of education, is dropping down. AI-powered assistants and conversational technologies are abundant in 2025. Because information is becoming ubiquitous, teachers employ technology to help students in improving their reasoning and critical thinking skills. Such technology is not intended to provide students with additional information, but helps to identify potential reasoning gaps, suggesting areas and resources that demonstrate different points of view.
Nevertheless, the increasing use of AI algorithms and big data does not come without risks relating to gender, class and racial disparities. AI often encompasses bias within its own design and implementation as shown in the following paragraphs regarding a recent product (classroom robot) that was biased and had to be withdrawn.
A biased classroom robot
Classroom robots was an AI-powered technology that observed classrooms (as a group of students), and that had to be withdrawn only a year after initial use due to accusations of algorithmic bias (attributed mainly to the design of the learning algorithms, that did not safeguard against the sue of biased training data). These risks arose from the design and implementation supporting the machine learning that took place in the training phase of the AI-based software, and even more so when the software was, in turn, used for teaching support. The classroom robots were meant to observe classroom activity as a whole and identify ways to improve collective learning. However, based on the training data used by these robots the suggestions were identified as biased towards specific content, e.g. if the data set used to draw conclusions from has been collected from a sample that is not free of bias, then the conclusions drawn will not be free of bias, and this will propagate onto the learning process, resulting in a biased AI-product.
Given the failure of this first attempt at deploying AI in classrooms, guidelines regarding future design of such software were put in place that included aspects such as:
how to consider bias when selecting training data,
how to balance transparency against performance issues during implementation, and
how to plan against bias in data sets that are collected and used dynamically, i.e. without human intervention (e.g. IOT).
Ongoing attempts to create appropriate ‘technology in education’ related policies across Europe aim to safeguard against similar biases. Although there exist deployments of AI in education in many schools both with student and teacher aids, safeguards against flawed implementations to avoid issues of bias in algorithms are important. Bias mostly arose because of reuse of past AI algorithms and software designs, especially in terms of diversity, an issue only recently improved in the area of AI algorithms.
Applications of AI in education in 2025
Education has not had significant budgetary growth in the past few years, especially as compared to the growth in the technology and business sectors. Yet now, new learning technologies are suddenly available in schools, as predicted by many technologists who have been promoting the integration of technology in education.
To appreciate the significance of these developments, the stalemate that the education sector experienced for a few decades prior to this change must be understood. The World Economic Forum in Davos in 2018 identified budgetary shortages and lack of innovation in the education sector, but it still took seven years (2025) to be able to confirm that new learning technologies have been introduced in schools. The World Economic Forum discussed the effects of emerging technologies, especially AI, and the
“imminent displacement of workers brought on by automation”, and added that “there is little doubt that … [education] has severely fallen behind the business world in realising the potential of new technologies – we need to shift our educational mindset to ensure that our children develop skills that can’t be replaced by a robot”.
Can the use of AI and big data be the tool needed to shift the current educational mindset? In 2025, AI is already present in the education field, through automated assessment feedback and virtual reality and augmented reality spaces. Whilst those have made an impression, they have not managed to motivate new standards and practices in the field, mainly due to the recent change, of introducing AI in schools to support learning, especially in the lifelonglearningcompanions for elementary school students.
Schools in Europe have been using a new in-class robot to observe class dynamics and learning trends, in order to identify group characteristics and support the teacher with delivery methods and content. The in-class robot features sociable skills: it has different facial expressions, head positions and tones of voice, which make it similar to a humanoid. Such features help the robot take over the role of teaching assistant.
Teachers are using AI technology in various ways in 2025, one of which is to automate grading of multiple-choice materials as well as more complex types of assessments. For essays or problem-solving assignments, the AI-generated marks are often matched to a human assessment marker as a verification tool.
Some university classes have hundreds of students. AI-powered robots help overworked professors to answer thousands of questions over the course of the semester. These robots can answer any curriculum-related question over an online classroom space without the students ever realising that they were talking with a robot. Similar robots are used in online classrooms to provide teaching support and student feedback to thousands of online students at a time. Student feedback verifies that the experience is not different from interacting with the class professor over the virtual classroom space.
Some students in 2025 have the opportunity to use an AI lifelong learning companion, as the learning buddies are entering many European classrooms. The AI-powered learning companions adapt to each student’s individual strengths and weaknesses in an effort to provide learning assistance throughout their student’s life. AI-powered learning companions help students with special needs by adapting materials to lead them to success, as well as providing personalised tutoringfor students outside of the classroom. When students need to reinforce skills or master ideas before an assessment, the AI-powered learning companion provides students with the additional tools they need, like revision lists based on their personal learning style or study guides organised according to the students’ preferences. Classroom robots learn to identify classroom weaknesses, such as when groups of students miss certain questions, and inform the teacher when material needs to be retaught. In this way, AI can also hold teachers accountable and strengthen best teaching practices. Nowadays, in 2025, these AI learning companions are a reality in many European countries, trained often by observing the young students and conversing with them.
The more data the robot collects over time, the more accurate the student profile becomes, and the more helpful the robot’s support.
The classroom as a concept is becoming redefined. This is a result of learning technologies making the classroom boundaries permeable to the learning content, but also because these technologies are becoming increasingly social, helping students to form learning communities based on their interests and skills. The grading system has also spanned outside of the classroom to the subject- and domain-specific learning communities. Teachers who develop content for such learning communities start to collaborate across national borders. Basic community courseware already utilize AI-powered facilitation, while advanced courses are requiring human facilitators to intervene online and offline.
Click here to watch making of B-AIM: