Artificial Intelligence Technologies in Education: Benefits, Challenges and Strategies of Implementation
AArtificial Intelligence Technologies in Education:Benefits, Challenges and Strategies ofImplementation
Mieczysław L. Owoc − − − ,Agnieszka Sawicka − − − , andPaweł Weichbroth − − − Wrocław University of Economics and Business,Komandorska 118/120 street, 53-345 Wrocław, Poland Gdańsk University of Technology,Faculty of Electronics, Telecommunications and Informatics,Department of Software Engineering,11/12 Gabriela Narutowicza Street,80-233 Gdańsk, Poland [email protected]://pg.edu.pl/en/home
Abstract.
Since the education sector is associated with highly dynamicbusiness environments which are controlled and maintained by informa-tion systems, recent technological advancements and the increasing paceof adopting artificial intelligence (AI) technologies constitute a need toidentify and analyze the issues regarding their implementation in edu-cation sector. However, a study of the contemporary literature reveledthat relatively little research has been undertaken in this area. To fillthis void, we have identified the benefits and challenges of implementingartificial intelligence in the education sector, preceded by a short dis-cussion on the concepts of AI and its evolution over time. Moreover, wehave also reviewed modern AI technologies for learners and educators,currently available on the software market, evaluating their usefulness.Last but not least, we have developed a strategy implementation model,described by a five-stage, generic process, along with the correspondingconfiguration guide. To verify and validate their design, we separatelydeveloped three implementation strategies for three different higher edu-cation organizations. We believe that the obtained results will contributeto better understanding the specificities of AI systems, services and tools,and afterwards pave a smooth way in their implementation.
Keywords:
Artificial intelligence · Benefit · Challenge · Strategy · Im-plementation. a r X i v : . [ c s . C Y ] F e b M. L. Owoc et al.
Research into artificial intelligence (AI) has tended to increase for a number ofreasons. The desire and excitement to create intelligent systems was the case inthe past in the academic community, while now it seems inevitable that theirdevelopment and inception have been addressed by the majority of organizations,representing almost all business sectors. It is not only the effect of the far-reaching technological progress such as microprocessors, data storage and globalnetworking, but also the impact of changes in business strategies. While thedebate on how AI will change business is at the top of the present-day agenda[1], education is already being challenged to reconceptualize existing teachingand learning methods by putting AI techniques and tools into service [2]–[4].Indeed, Sameer Maskey, founder and CEO at Fusemachines, an AI Educa-tion and AI Talent Solution provider based in New York City (USA), in aninterview with Forbes magazine published on 8th June 2020, said that [5]: “Itwill be important for educators and policymakers to explore the intersection ofeducation and artificial intelligence. The application of machines in learning en-vironments is only one variable in a multifaceted equation. We have to considerbarriers that prevent an even distribution in technological resources and how toovercome them. We must also ensure that teachers are prepared and empoweredto leverage artificial intelligence. Assuming these elements are addressed, thepossibilities of AI-powered learning are infinite”.Inspired by his words, and apart from the financial costs and profits, thequestion arises what are the specific benefits and challenges of implementingartificial intelligence in education? Since little work has focused on this area, thefirst research goal is to identify and analyze its possible benefits and precedingchallenges (see Sec. 2.2 and 2.3, respectively). In this case, a qualitative genericthematic analysis was undertaken due to its flexibility to collect descriptive datain the narrative form.Our study is also driven by a concern for the implications of the humanfactor due to the increasing evidence of the apprehension raised against theproliferation of artificial intelligence since the variety of its applications seems tobe immeasurable, ranging from computer games [6,7], decision-making [8]–[10],education [11]–[16], enterprise management [17]–[19], grid computing [20]–[23],knowledge management [24]–[27], learning systems [28]–[31], ontologies [32]–[34],smart cities [35]–[37], and software engineering [38]–[40], to name just a few. Onthe contrary, we attempt to argue that AI can amplify educational effectiveness,concerned with sharing, developing and disseminating knowledge, at the sametime preserving human autonomy, agency and capabilities. Therefore, the secondresearch goal is to analyze and evaluate the usefulness of existing AI technologies(see Sec. 2.4).To explore these two goals, we collect, analyze and review a plethora ofinformation sources. Their addresses were obtained from the Google web searchengine and Google Scholar website by using combinations of keywords such as:“artificial intelligence”, “adoption”, “acceptance”, “advantages”, “disadvantages”, rtificial Intelligence Technologies in Education: (...) 3 “implementation”, “deployment”, “education”, “benefits”, “challenges”, “risks”, and“technologies”.What is much rarer, however, was to find applicable and relevant implemen-tation strategies. Nevertheless, based on our previous results, which provide solidtheoretical foundation, we also designed and created a generic strategy, able tobe applied in a strategic plan regarding any of the AI systems, services or tools inall industries, and in organizations of a variety of sizes. In particular, our strat-egy addresses the “what” and “why” of the activities, embedded in a five-phaseprocess model (see Sec. 3).To verify and validate its design, we separately developed three implementa-tion strategies (see Sec. 4), for three different non-public higher education insti-tutions, namely the “Copernicus” Wroclaw Computer Science and ManagementUniversity (WSZI), WSB University in Gdansk (WSB) and Jan WyżykowskiUniversity (UJW). By adopting a qualitative study design, as an exploratory,descriptive approach, in the first step we collected all of the necessary data, usinga specific thematic approach, while in the second step we analyzed the data ina fashion reflecting the aim of recognizing circumstances and challenges relatedto the subject matter.It is worth noting here that nowadays, non-public universities play an essen-tial role in higher education [41]–[43] and have to compete very hard in orderto gather potential students [44]–[46]. Their position increasingly depends onthe quality of the education and the managerial competencies of the universitygovernance [47]–[49]. In both cases, applying intelligent technologies seems to bea must if one considers their competitiveness and development. Yet, the level ofits implementation is still relatively low in comparison with the business sector.On the other hand, there are a few cases documented, giving an idea in whichareas AI methods have been implemented within higher education institutions(HEIs) [50]–[52]. In particular, intelligent technologies are gradually being im-plemented in non-public universities [53], usually being part of the strategy thatsets up a framework of priorities [54]. However, to the best of our knowledge,very few studies have considered the benefits and challenges affecting the imple-mentation of AI technologies within university emerging set-ups.The rest of the paper is structured as follows. Section 2 discusses artificial in-telligence (AI) in the education sector, introducing the basic concepts, benefitsand challenges related to its implementation, as well as a discussion of mod-ern AI systems and tools. Section 3 presents a strategy implementation model,conceptualized by a five-stage generic process along with the corresponding con-figuration guide. Section 4 describes and analyzes the case studies, illustratinghow AI implementations, based on underlying decision processes, are conductedin practice. Finally, we conclude in Section 5.
By design, intelligent technology is a method which uses knowledge to achieveconcrete purpose in efficiency. At present, there are the following intelligent tech-
M. L. Owoc et al. nologies: multi-agent, machine learning, ontology, semantic and knowledge grid,autonomic computing, cognitive informatics and neural computing. The promptadvances in these fields have already brought substantial changes in education,opening up new opportunities and challenges to teach and learn anytime andanywhere by providing new methods and systems that aim to stimulate innova-tive teaching and ultimately improve learning outcomes.
The continuous progress of modern information technologies is strictly connectedwith the presence of implemented artificial intelligence techniques. During theover 60 years of development of artificial intelligence, several intelligent ap-proaches have appeared in almost all sectors of modern life. Therefore, one cantalk about the new generation of AI, including the potential power of the currentsolutions and the variety of applied techniques. The crucial components of suchan understanding of AI 3.0 are presented in Figure 1.
Fig. 1.
Artificial Intelligence 3.0 [55].
Particular categories of AI can be combined in the final applications; someof them seem to be obligatory (knowledge, reasoning, processing) while othersare employed for the specific solutions where knowledge should be permanentlyupdated (machine learning) or requires cooperation between specialized agents(multi-agent systems). Either way, the current solutions are not quite satisfactory– the level of the obtained progress is less than human intelligence. The nextimaginable stages are still before us – we are heading toward Artificial SuperIntelligence passing in the meantime through Artificial General Intelligence (seeFig. 2). rtificial Intelligence Technologies in Education: (...) 5
Fig. 2.
Future evolution of Artificial Intelligence [56].
To sum up, the landscape of Artificial Intelligence in terms of using themain categories is rather stable. Yet, learning-based techniques are playing anincreasingly significant role. However, the specialty of the application areas candetermine the shape of implementation of artificial intelligence methods. The ed-ucational sector and the specifics of non-public universities are real determinantsin defining the implementation of intelligent technologies.
With the advent of AI in the mid-1950s to the present day, the proliferation ofits methods and techniques has made it possible to develop intelligent systemswhich are increasingly relevant in education and training. For instance, Nuance,the high-tech company from Burlington (Massachusetts) [57], has implementedspeech recognition software that can be used both by students and faculty. Theapplication can transcribe up to 160 words per minute and is particularly usefulfor students who have limited mobility or struggle with writing. The availablefeatures also enhance word recognition and spelling. Teachers can apply thesoftware to prepare homework and assemble and schedule recurrent tasks suchas sending notifications or emails.Another company, Knewton, is promoting its newest product, Alta [58], as acomplete courseware solution that combines expertly designed adaptive learningtechnology with high quality openly available content. In other words, Alta helpsidentify drawbacks in a student’s knowledge by providing relevant coursework,and supports teaching activities at different educational levels.Cognii is another provider of artificial intelligence by virtue of virtual as-sistants (VAs) that combine the powers of conversational pedagogy with theconversational AI technology [59]. Interestingly, the open-format applied to the
M. L. Owoc et al.
VAs’ responses is claimed to improve critical-thinking skills. The VAs also pro-vide real-time feedback, and individual tutoring, customized to the particularstudent’s requirements.Querium, a successful start-up from Austin (Texas) [60], helps students mas-ter critical STEM skills by delivering a customizable STEM tutoring program ofpersonalized and effective lessons which works on desktop computers and smart-phones. Querium’s AI provides teachers with insights into the student’s learninghabits and highlights areas in which the student should improve.Century is another successful start-up, established in 2013 in London byPriya Lakhani [61]. Behind its success is a diverse team of teachers, technol-ogists, neuroscientists and parents. Their platform utilizes data analytics andcognitive neuroscience to create personalized learning plans and reduce work-loads for teachers. Moreover, the AI platform tracks student progress, identifiesknowledge gaps and delivers personal learning recommendations and feedback.The teacher dashboard allows them to monitor individual student and whole-class performance.These examples show only some of the possible applications of AI in theeducation sector. Nevertheless, we argue that the above allows us to claim that ata general level, the constellation of AI methods and tools leverage both learningand teaching. With time constraints, limited resources and the abundance ofnew, incoming knowledge, the engagement of artificial intelligence seems to bea must to deliver a competitive environment, facilitating both the learning andteaching processes.According to the latest literature concerning the application of artificial intel-ligence in the education sector, computer-assisted tutoring represents the majorapplied field of AI [62]. At present, we can observe that the rate of develop-ment of intelligent educational software (IES) has increased due to the high anddynamic students demands [63].The main problems are related to content flexibility and adaptability, andfor reusability, sharing and collaborative development of the learning objectsand structures [64]. Moreover, providing user-oriented content depends on threefactors, namely the domain model, user model and instructional task model [65].
The advantages of artificial intelligence applications in education are vast andvaried. Here, everything can be considered to be beneficial if we are thinking ofanything, for example a computer program, that can efficiently perform any taskthat would normally rely on the intelligence of a human. Based on the state-of-the art research in this area, we outline nine areas in which AI methods canbring added value for both learning and teaching activities [66].The first benefit concerns automated grading which simulates the behaviorof a teacher to assign grades to the answer sheets submitted by the students.It can assess their knowledge by processing and analyzing their answers, givingfeedback and recommending personalized teaching plans. rtificial Intelligence Technologies in Education: (...) 7
Secondly, intermediate spaced repetition aims at knowledge revision whensomeone is just about to forget. It is worth noting that Polish inventor PeterWozniak [67] introduced the SuperMemo application, which is based on the effectof spaced repetition. The app keeps track of what a user is learning, and whenhe/she is doing it. By applying AI techniques, the application can discover whena user is most likely about to forget something and recommend revising it.Thirdly, feedback loops for teachers, aided by machine learning and naturallanguage processing techniques, improves the quality of student evaluations. Forexample, a chatbot can collect opinions via a dialog interface similarly to a realinterviewer but with a small amount of work required by the user. Moreover, eachconversation can be adapted according to the student’s personality and providedanswers. A chatbot can even formulate the reasons for particular opinions.Fourthly, to support teachers in their classroom work, one can put intouse virtual facilitators. For instance, at the Georgia Institute of Technology onKnowledge-Based Artificial Intelligence (KBAI) class, students were introducedto a new teacher’s assistant named Jill Watson (JW) [68], who has been operat-ing on the online discussion forums of different offerings of the KBAI class sinceSpring 2016. JW autonomously responded to student introductions, answeredroutine, frequently asked questions, and posted announcements on a weekly ba-sis. In the fifth place, Watts introduced chat campus based on the IBM Watsoncognitive computing technologies [69]. In brief, students at Deakin Universityhave asked IBM Watson 1600 questions a week to learn the ins and outs oflife on campus and studying in the cloud. Within 12 months of implementingWatson, due to the enhanced quality of the student know-how at Deakin, thisground-breaking solution has handled more than 55,000 questions from students.Furthermore, the school is progressing its use of Watson, broadening its capa-bilities and teaching the system to understand new sources of information.Personalized learning is the sixth example of AI applications in the educationsector. In general, it refers to a variety of educational programs in which thepace of learning and the instructional approach are customized and eventuallyoptimized for the needs of each learner [70]. In particular, the content is tailoredto the learning preferences and specific interests of each student.The seventh example—one of the most promising—is adaptive learning (AL).While the traditional model of classroom education, continues to be very muchone-size-fits-all, on the contrary, AI-powered AL systems are designed to opti-mize learning efficiency. For example, Yixue Squirrel AI (Yixue) collects andanalyses students’ behavior data, updates learner profiles, then accordingly pro-vides timely individualized feedback to each student [71].Since cheating is a concern for all teachers, AI-powered anti-cheating systemshave been presented as another (eighth) application of AI in the education sec-tor. Proctoring is software which secures the authenticity of the test taker andprevent him/her from cheating as a proctor is always present during the test[72].
M. L. Owoc et al.
The last solution argued by Watts is data accumulation and personalization.For instance, learning grammatical rules can be aided by examples only fromthe domain being the subject of personal interest [73].
There are several approaches to planning and organizing the implementation ofAI methods in the education domain [74]–[81], but discussion about the essentialchallenges for decision-makers is still ongoing. To the best of our knowledge,the list of potential challenges that influence the implementation of intelligenttechnologies concern: – strategy refers to a general plan of implementation to achieve one or morespecific long-term goals accordingly to a schedule established and agreedwith all interested stakeholders; – organizational maturity refers to its employees, processes and technologyreadiness and capability with respect to the adoption of artificial intelligencetechnologies; – data governance , refers to data principles, quality, meta-data, access re-quirements and data life cycle; since machines learn on the basis of data, datagovernance is a crucial facet of the implementation and further maintenanceof AI; – infrastructure , being the combination of hardware and software systems,is particularly acute due to compatibility and integration issues.As one would expect, it is important to establish a strategy that defines the goalswith regard to the AI implementation and provides a means to manage them.The strategy itself might take the form of a mix of qualitative and quantitativeapproaches. The former aims to describe how the goals will be fulfilled, whilethe latter aims to decide if the goals are fulfilled and which goals are fulfilled.The fulfilment of the goals cab be expressed in quantitative numbers, or/and inqualitative terms.In general, maturity is a synonym of “full development” or “perfected condi-tion,” and since any organization is a living entity, it grows over a period of timeand learn from its decisions and outcomes. Therefore, all organizations seem tobe at some stage of maturity, striving forward to development and perfection.From a strategic point of view, we stress the importance of the high level oforganizational maturity due to the changes spanning across core dimensions ofstrategic management such as: alignment, performance measurement and man-agement, process improvement and sustainability. In the context of our study,maturity assessment should encompass external and internal benchmarks, de-scribing the organization readiness and capability to adopt AI technologies.Another challenge is data governance, which is related to the system of dataorganization, collection, control, storage, usage, archival and destruction. Thepath of setting up data governance is driven by a specific program, supported byparticular policies and procedures, and communicated by organizational lead-ership and management. In general, the regulations must provide all of the rtificial Intelligence Technologies in Education: (...) 9 necessary means to preserve the following generic requirements: accessibility,availability, completeness, accuracy, integrity, consistency, auditability, and se-curity.The last, but not least significant, challenge concerns the infrastructure whichencompasses all of the hardware installations and the software entities. Recentadvancements in the artificial intelligence technology landscape have introducedspecific requirements toward hardware capacity and software capabilities. In aneffort to integrate these cutting-edge technologies with the existing systems, onehas to incorporate solutions that underpin a flexible and scalable end-to-end in-tegration. Enabling on-the-fly software asset configurations and reconfigurations(in the case of enabling/disabling particular services) facilitates an “assembly-from-parts” model for implementing new and updating existing AI applicationsfrom a catalog of services.We argue that the above-mentioned challenges are essential to take into con-sideration while preparing the discussed scenarios (see Section 4). Our researchin essence attempts to identify and analyze the issues related to implementingAI in education, however it still lacks the empirical evidence which might atleast confirm our perceptions of the studied phenomenon.
Undoubtedly, artificial intelligence has had significant influence on various in-dustries, leveraging their effectiveness, productivity and profitability. This alsoapplies to the education sector which has been committed to several reforms,addressing key sectoral issues and including incorporating artificial intelligencetools and methods.It is believed that such a shift will bring a new perspective to many facetsof existing learning and teaching techniques. Toumi, in his report from 2018,claims that [67]: “(. . . ) in the domain of educational policy, it is important foreducators and policymakers to understand AI in the broader context of thefuture of learning. (. . . ) As AI will be used to automate productive processes,we may need to reinvent current educational institutions. (. . . ).”This opinion and many others thereafter argue the need to better compre-hend the impact of artificial intelligence on education. However, first we need todevelop a model that will drive the process of implementation and result in thesmooth deployment of AI systems. At this point, we argue that the below modelis suitable under general contexts, and is thus applicable to any organization.The five-stage process shown in Figure 3 organizes, arranges and systematizesthe tasks into consequent groups. The particular tasks employed in each stagemight cross other phases, to an extent which depends on the context of the imple-mentation (e.g. strategy, organizational maturity, data governance). Therefore,the five stages are interdependent with each other, whereas the duration andlabor intensity may significantly vary. The model of the implementation processconsists of five stages:
Fig. 3.
General strategy implementation model. Plan and analyze stage concerns all activities associated with the creationand maintenance of a plan which describes a list of steps with details of thetiming and resources required to achieve the desired goals, along with thebudget and general time frames.2.
Design and specify stage aims to prepare and establish the structureand organization of the system as well as to define the functional and non-functional requirements. In other words, a specification should address allenumerated goals in the first stage.3.
Implement and configure stage can be interpreted twofold; the systemimplementation is the process of creating its source code by the softwaredevelopers, or the system implementation is installing and configuring thesoftware applications.4.
Test and evaluate stage aims to ensure that the actual software is freeof defects, and to check and determine whether it matches the expectedrequirements, as defined in the second phase.5.
Monitor and support stage concerns the surveillance process on measur-able events regarding the performance of the system, as well as providingassistance and help to the users of the system required due to any issues andincidents encountered.Moreover, it should be emphasized that it is necessary to identify the specifici-ties of AI software and evaluate existing resources, including hardware capacity,software compatibility and IT personnel. However, anticipating all of the relevantresources is only possible under exceptional circumstances exceptionally possibledue to uncertainty regarding especially the human factor and data quality (usedto teach AI algorithms).
One should keep in mind that the strategy is a “how” the goals (objectives)will be achieved by the means (resources). In the context of this study, thestrategy consists of the desired objectives and corresponding actions in order toimplement an AI solution. Having said that, the structure of the implementationconfiguration is as follows: – general purpose and vision, – scope, specific objectives with priorities and within time frames, – reference to the procedures used, regarding all five stages of the implemen-tation model (see Fig. 3), rtificial Intelligence Technologies in Education: (...) 11 – reference to the AI systems, services and tools which are existing and ableto be included in the strategic plan, or in the case of developing a newsystem, a description of the possible software vendors, trusted advisors andconsultancy agencies, and individuals experts, all with designated roles andresponsibilities.This perspective conceives the know-how as a highly deliberate, descriptive,and logical one, involving a sequential, rational and analytical collaborative workof all interested stakeholders. The configuration strategy is conceptually aimedat achieving the master plan by precisely following the implementation process.However, this form of strategy configuration can be mostly applied in organi-zations which at least manage projects in accordance with a policy that aims toidentify and monitor the progress towards project performance objectives (leveltwo). Respecting the autonomy of higher education institutions, we would expect veryindividual approaches to deploying a single or suite of AI techniques for theuniversity. The selected non-public universities that participated in the researchare different in terms of geneses, formal statements as well as levels of computerinfrastructure and areas of education. Before presenting the proposals for imple-menting artificial intelligence solutions in the inquired HEIs, short descriptionsof the universities are delivered, focusing on their areas and levels of education.The state-of-the-art regarding the maturity of the implemented IT solutions arealso discussed.
The university which is the subject of research ofthis article, namely The "Copernicus" University of Information Technology andManagement, has been operating on the educational services market for nearly20 years and is one of 384 non-public universities in Poland.Since 2001, it has educated 4,000 graduates, mainly in the area of new tech-nologies. Further observations and conclusions presented in the article will bebased only on their own observations of business practice and the environmentand competition due to the lack of data from external actors evaluating thequality, innovation and conditions for studying and living for students. Start-ing from 2019, the university has been conducting studies solely in the field ofComputer Science in a full-time and part-time system at two levels, includingpost-graduate studies and courses.The organizational structure of the university is a simplified structure withone department and full decision-making ability of the Chancellor, who is alsothe founder. His decisions are binding on financial, program and student mat-ters. For the analysis of the organizational structure, one can state that WSIZ "Copernicus" omits the very important role of Rector. Usually, the rector in anon-public university is the person who deals with the organization of educationand training, but his decisions are not binding without the authorization of theChancellor. In addition, the described university is a unit educating a total ofabout 700 students in a full-time and part-time systems.Among others: the department of the Rector’s office, where there are toofew employees and students. It is worth noting that if there were changes in thefunctioning of the organizational structure, it could be possible to reorganize thework and introduce intelligent technologies into these structures, introducing theinnovation which would lead to the beginning of changes.The financing of a non-public university, which clearly affects the develop-ment and functioning of educational institutions, departs from the possibility offinancing public entities. Obtaining funds or research grants is difficult due tothe small number of students.There is also no obligation to conduct research, which makes it more difficultto win in competitions. All retrofitting of the university halls are covered onlyfrom the profit earned, which consists of tuition payment and fees for programdifferences, conditions. When it comes to the issue of payment for tuition, thestudent has the privilege of financing science on the basis of a scholarships ob-tained, which are largely similar in terms of the procedures for obtaining fundsin public universities.A non-public university, even though it is a university, is also an enterprise.And just as a company will not exist without clients, a private university willnot be able to break out of the educational market without the influence andopinion of its potential future students.If we focus our conclusions on the differences in the functioning of these twotypes of universities, we should also mention the change in the law on highereducation, which imposed student limits per lecturer on public universities. InWSIZ "Copernicus", the current number of students per lecturer is 25–30 stu-dents, which definitely departs from the provisions of law for public entities.In addition, due to the lack of sufficiently qualified teaching staff and theChancellor’s desire to reduce employee costs, the university introduced the pos-sibility of writing diploma theses to teachers with a master’s degree, providedthat formal care over diploma theses will be carried out by a promoter with adoctorate degree. However, he leads the student through the whole process ofwriting the work, advises him, gives his opinion on his progress and is listed inthe diploma documentation. The content supervisor, who is subject to the pro-moter, is a person with a doctorate degree or above. Such a solution is allowedby the applicable law and applied in WSIZ "Copernicus", however, there is noinformation whether and to what extent public HEIs use it.Looking at the above information about the structure, method of financingand processes taking place in the area of education and work organization ofthe described non-public university, it can be stated that although it is an entityeducating students similarly to state universities, it is significantly different fromthem. rtificial Intelligence Technologies in Education: (...) 13
Obviously, it is not disputed that issues such as legal regulations, the shapeand form of the graduate and student documents, requirements regarding ECTSpoints, and reporting or deadlines, are almost identical. However, they are soindividual in their approach to the quality of education and the candidate thateach of them should be considered separately. In the same way, when designingand implementing business analytics solutions, it should be taken into accountthat what works in a given non-public unit does not necessarily have to serveanother organization, especially a state one.
Case 2. WSB University in Gdansk
The WSB Universities is the largestgroup of university schools of business in Poland. All of them are found in toppositions in rankings of higher education institutions. Founded in 1994 by TEBAkademia, WSB launched its first field of study: “finance and banking”. Presently,the WSB Universities are located in 10 cities across Poland: Gdańsk and Gdy-nia, Bydgoszcz and Toruń, Wrocław, Chorzów, Opole, Poznań, Szczecin, andWarszawa.The organization of this particular University is typical for HEIs in Poland.The faculties and departments are oriented on educational sectors and there aremany supporting divisions mentioned in the characteristics of the previous Uni-versity, but particular units cooperate strongly, including by utilizing commonplatforms (webpages) for students and academic staff.The WSB educational offer covers both bachelor’s and master’s programs.The duration of the former is 3 years (6 semesters) while the latter is 2 yearslong. It is worth noting here that a student can pursue second-cycle programsin a field that is not directly related to that of their first-cycle degree. In thatway, one can extend their area of professional expertise and hence enhance theiremployability. Each program is focused on developing practical competenciesand soft skills.Each university is governed by their own statutes and regulations, but areintegral to the make-up of the ownership authorities. Moreover, each individualuniversity has its own internal procedures, regulations, and local authorities.The university is governed by the rector along with the chancellor. At the headof each faculty are the dean and vice-deans, supported by the proxies.University Faculties organize teaching and research into individual domains,or groups of subjects. Their work is normally organized into subdivisions calleddepartments. The University’s administrative and support departments supportthe running of the University and contribute to research, teaching and interna-tional cooperation with other universities worldwide.The organization claims to maintain a high performance culture that is in-spirational and motivating, providing internal and external funds to support fac-ulty staff development. The aim of each subsidiary is to provide an attractive,sustainable, vibrant, and accessible campus, upheld by a contemporary virtualenvironment in which students, staff, and engaged stakeholders can interact, andshare information and knowledge.
Case 3. Jan Wyżykowski University in Polkowice.
Jan Wyżykowski Uni-versity was created thanks to the cooperation of Polkowice Commune and Polkow-ice County. The founder of the University is the ZAMPOL company, whichPolkowice Commune and Polkowice County own all of the shares of. Currently,UJW runs studies within the Bachelor’s, Engineer’s, Master’s programmes andpostgraduate studies and courses of Education offerings: Bachelor’s studies (3-year studies) in the fields of: Administration, Pedagogy and Management. Engi-neer’s studies (3.5-year studies) in the fields of: Information Technology, Mecha-tronics, Logistics, Mining and Geology, Production Management and Engineer-ing. Master’s studies in the fields of Management (2-year), Mechatronics (1.5-year) and a uniform (5-year studies Law) are the newest in the offer.All educational majors are offered as extramural studies, so basically, stu-dents gain practical experience working for different companies. The second im-portant feature of the university is that it acts in a very specific region knownfrom its copper industry. Most of the students are employed in companies con-nected with this industry so their professional expectations are strictly tied withcopper infrastructure or reflect more general positions typical for medium-sizedcities (administration and management).Nowadays, the information systems present in UJW support typical trans-action processes: student enrollment, planning and evidence of classes and somefunctions offering by online education. It is quite a well-functioning universityinformation systems but without functionality specific for AI methods.
The solutions in Figure 4 use the latest technologies, which confirms that manyrepetitive tasks and procedures could benefit from the support of AI systems,thus offering new development opportunities and fields of study for higher edu-cation.It is important to remember that there are many often more complex andnon-standard procedures at universities. The processes of obtaining the intendedeffects from the implementation of the proposed solutions based on AI should bein accordance with applicable standards and the law regarding higher educationand also take into account the capabilities of the university. Each replacementor support through the implementation of technological innovations is intendedto shorten the time of document circulation, reduce administrative costs andimprove the quality of education, which is certainly worthy of attention.Naturally, introducing such a solution based on new technology solutions isnot without its own inherent risks. Non-public universities would need to exerciseextreme caution in protecting students’ personal data and would need some levelof human oversight to monitor every AI method.
Case 1. WSIZ university
Building creative solutions into the work of a non-public university is also helpful in creating new products and services. The im-plementation of smart technology solutions at a private university will foster the rtificial Intelligence Technologies in Education: (...) 15
Fig. 4.
Scenarios of implementation of selected AI methods for WSIZ, WSB and UJW.6 M. L. Owoc et al. building of scientific and educational progress, which will translate into satis-factory financial results and will strengthen the position on the non-public uni-versity market and introduce the element of innovation to the current activitiesof the unit. As part of the implementation work envisaged at the "Coperni-cus" University of Information Technology and Management in Wroclaw, it isplanned to create a chatbot—an intelligent application for managing responsesand relations with students and candidates for studies.This solution, in connection with the personnel problems of the non-publicuniversity described would answer simple questions about the dates of the ses-sions, exams, classes and inform about the recruitment schedule and the re-cruitment documents needed. As part of the planned implementation, it is alsoplanned to reorganize the signing of learning contracts, which would consist inenabling their signing in electronic form via a link sent, generated by the internaluniversity system, which is consistent with the provision in art. 60 cc. Supportfor such a contract would take place in the following steps:1. A link would be sent to the student’s e-mail address to the page about thelearning contract with a request to read its contents.2. After learning the terms of the contract, the student would be asked to acceptit by clicking the “I accept the terms of the learning contract” button.3. After choosing this option, the contract is concluded electronically.Afterwards,the student accepts the presented conditions, the confirmation along with thecontract is sent automatically in a pdf file to the dean’s employee. It is thenprinted and archived to the student’s personal file.In addition, the plans of AI implementation in WSIZ also include the prepa-ration and full implementation of a voice guide for students and candidates withimpaired vision. The voice guide presents an educational offer and discusses pay-ment, calls the phone number for the department, and checks the plan for thatacademic year.The introduction of intelligent solutions for the work of the "Copernicus"University of Information Technology and Management in Wroclaw creates op-portunities to facilitate procedures in the recruitment process, and reduce theuse of office machines, which also affects the improvement of the environment.It relieves the work of the dean’s office, which is the most important area incontact with students and candidates.Of course, apart from the positive aspects, one may ask the question whetherintelligent technologies strive to improve human work and make it easier forthem to complete certain processes to become more attractive on the market,or whether it is simply interference in human work and increasing unemploy-ment. However, believing only in the good aspects of the case, we believe thatthe solutions and techniques of business intelligence will significantly improvethe quality of the education process and assist in acquiring new students at the"Copernicus" University of Information Technology and Management in Wro-claw. rtificial Intelligence Technologies in Education: (...) 17
Case 2. WSB Universities
We have identified three administrative areaswhich are planned to be fully or partially automated by implementing artificialintelligence tools. The list is given below (not in order of priority).1.
Grouping, sorting and responding to emails . From our employees, weknow that replying to emails is a time-consuming job. Moreover, repetitiveemail conversations are also frustrating and can be demotivating in the long-term. There are a number of tools which are being considered to be includedin a pilot study, namely: AI Email Smart Answer [82], OMQ Reply [83], andNotion [84]. All of these tools are able to automate responses to emails andeventually replace the manual work of employees. Recurrent students’ re-quests are automatically recognized and answered by the system. Obviously,the list of features of the AI-powered solutions doesn’t end here.2.
Scheduling appointments . Scheduling meetings with multiple studentsis a labor-intensive task. We believe that instead of manually respondingvia email to schedule an appointment and check the calendars of everyoneinvolved, we could implement intelligent agents that detect and recognizecertain phases in incoming emails, eventually proposing appointment timesaccording to individual availability, and schedule appointments based on theattendees’ responses. Again, we have selected a few tools which are plannedto be tested in a pilot study, namely: Julie Desk [85], ArtiBot [86], and Hen-drix.ai [87]. Meetings are still a crucial method of organizing and planningwork, but they are a waste of time unless one accurately captures what wasdiscussed and agreed upon. By design, the agents require access to individualcalendars, email accounts, social media profiles and location data to providethe necessary data to the AI-based inference engine. In return, they simplyserve as virtual assistants, capable of preparing meetings, dialing into con-ference calls, turning on a video projectors, loading presentations, removingoutdated and duplicate contacts, and many more otherwise tedious tasks.3.
Customer service AI chatbot . A chatbot is simply a software agent thatcan simulate a conversation with a user in natural language in a real timetrough messaging applications, websites, mobile applications, or even overthe telephone. The requests reported by the students dramatically increasebefore particular events, such as bachelor and master diploma exams, the be-ginning and the end of the academic year, as well as from the candidates whointend to become students. Therefore, customer service employees are regu-larly inundated with follow-up calls, support requests, frequently and repet-itively asked questions, confirmation emails, complaints, and many more. Toface these issues, IBM tells us that “chatbots can help businesses save oncustomer service costs by speeding up response times, freeing up agents formore challenging work, and answering up to 80% of routine questions” [88].Moreover, unlike live agents (employees), software agents don’t need lunchhours or coffee breaks, and are not absent due to holidays, or illness, or any othernatural disasters which can put human lives at risk. At the moment, our pilotstudy includes the following AI chatbot software to be tested: ActiveChat [89],
Respond.io [80], and ChatBot [91]. We expect to uncover sizeable advantages byimplementing chatbots across the organization.On the other hand, we are aware that it will not be an easy and effortlesstask. On the contrary, based on our learnings, we have identified and listed thefollowing three major challenges, namely: security and privacy, obstacles andburdens due to polish language complexity, data input for machine learningalgorithms. Undoubtedly, chatbots are one of the most promising enterprise AItechnologies, however, achieving maximum business value from them requiresfrom us extensive work and persevering determination.
Case 3. UJW University
Nowadays, the UJW educational unit is trying tobe competitive as a local experimental HEI. Therefore, the educational offeringrelates to regional needs especially strongly connected with the copper indus-try and other companies located in the region. UJW was a leader of Erasmusgrants devoted to improving the educational level through applying new teach-ing methods using AI (e.g. project DIMBI [92] and related [93]–[96]). Thereforeall potential solutions in this HEI are strictly connected with the results of theseprojects extended by discovered niches in education. Three directions of AI tech-niques to embrace are:1.
Using innovative methods of teaching selected courses . At the begin-ning, practical abilities and skills refer to the data science and mechatronicmajors.2.
Smart agents (SAs). SAs are used to automate administrative procedureswith the aim to simplify daily tasks.3.
Chatbots . Implementation of chatbots for communication with actual andpotential students. This project is comparable to that presented in the pre-vious universities.Educating staff for the constantly changing trends and creating innovationsboth in the area of education and the functioning of a private educational unittranslates into profit and brand visibility and the prestige of education. Thesearch for creative solutions is the equation between the known, acceptable orderand the chaos that innovation can bring. However, if innovation is treated asa course of action, as a way of managing, without worrying about disturbingrelative stability, after a period of time, it can be seen that it has marked outnew directions and possibilities of creating a competitive advantage.To sum up, we think that the automation of a universities’ administrativetasks and customization of their student-oriented activities are not only possible,but imminent. The goal of AI technologies is to make human-like judgmentsand perform tasks in order to downsize employees’ workloads. Today, academiccommunities are intensively developing and studying this field of AI applications.Indeed, the findings from research commissioned by the Microsoft show almostcomplete acceptance among educators that AI is important for their future—“99.4 percent said AI would be instrumental to their institution’s competitivenesswithin the next three years, with 15 per cent calling it a game-changer” [97]. rtificial Intelligence Technologies in Education: (...) 19
However, in this paper, including the assumptions of the first pilot study,we “only” considered selected administrative areas, performed by the employeesnot directly responsible for education. Nevertheless, we believe that AI has far-reaching potential to change the way of teaching and learning. Undeniably, theincoming shift has its advocates and opponents whose proofs and claims shouldbe always carefully judged.
As the technologies of artificial intelligence evolved, so did the domains andpractices of their implementation in education. Current trends have imposednew requirements on the organization and management of both teaching andlearning. There are three interrelated aspects of this, one of which arises fromthe recent advancements and innovations of the cutting-edge machine learningmethods and Internet-of-Things devices.Secondly, the focus of both teachers and learners concern very large vol-umes of information and knowledge resources, freely available on the Web. Theirgrowth in size and number seems to be an endless road to discover, but thereis more and more evidence to help us pick the right direction. Thirdly, cur-rently the competitiveness of higher educational institutions depends stronglyon the increase in the effectiveness of learning methods, strongly supported byAI technologies and tools [98]–[100].At a time when many education entities are stretched to capacity, and learn-ers experience long wait times for on-site counseling, AI solutions could providesome facilities. It is therefore recommended that organizations should use solu-tions that are supported by the latest technological solutions, in order to improvethe quality of education, and to minimize errors in the circulation of adminis-trative documentation and the course of study.To sum up, we argue that the role and impact of artificial intelligence hasincreased in the education and learning contexts. The academic sphere is becom-ing more effective and personalized on the one hand, as well as global, context-intensive (multi-cultural) and asynchronous on the other. The intersection ofthree areas, namely data, computation and education has set far-reaching con-sequences, raising fundamental question about the nature of teaching: what istaught, when it is taught, and how it is taught.
References
1. Bullock, L. The Top 6 Ways That Artificial Intelligence WillAffect Your Business In The Near Future. Forbes Maga-zine.
2. Hao, K. (2019). China has started a grand experiment in AI education. It couldreshape how the world learns. MIT Technology review. Luettu, 1, 2019.0 M. L. Owoc et al.3. Garrett, N., Beard, N., & Fiesler, C. (2020). More Than" If Time Allows" The Roleof Ethics in AI Education. In Proceedings of the AAAI/ACM Conference on AI,Ethics, and Society (pp. 272-278).4. Neller, T. W. (2016). AI education: Birds of a Feather. AI Matters, 2(4), 7-8.5. Forbes (2020). Artificial Intelligence In Education Transforma-tion.
58. Knewton (2020).
59. Cognii (2020).
60. Querium (2020). http://querium.com/
61. Century (2020). https://activechat.ai/?fp_ref=tooltester [data accessed:2020-06-20].90. Respond (2020). https://respond.io/?lmref=yCoaYw [data accessed: 2020-06-20].91. ChatBot (2020). [data accessed: 2020-06-20].92. Owoc M., & Pondel M. (2018): Selection of Free Software Useful in Business intel-ligence. Teaching Mathodology Perpsective.[in: E. Mercier-Laurent, D. Boulanger(eds.): Artificial Intelligence for Knowledge Management. Springer.93. Weichbroth, P., Owoc, M., & Pleszkun, M. (2012). Web user navigation patternsdiscovery from WWW server log files. In 2012 Federated Conference on ComputerScience and Information Systems (FedCSIS) (pp. 1171-1176). IEEE.94. Redlarski, K., & Weichbroth, P. (2016). Hard lessons learned: delivering usabilityin IT projects. In 2016 Federated Conference on Computer Science and InformationSystems (FedCSIS) (pp. 1379-1382). IEEE.95. Owoc, M., Weichbroth, P., & Żuralski, K. (2017). Towards better understandingof context-aware knowledge transformation. In 2017 Federated Conference on Com-puter Science and Information Systems (FedCSIS) (pp. 1123-1126). IEEE.96. Owoc, M. L., & Weichbroth, P. (2018). A Note on Knowledge Management Educa-tion: Towards Implementing Active Learning Methods. In IFIP International Work-shop on Artificial Intelligence for Knowledge Management (pp. 124-140). Springer,Cham.97. Ayoub, D. (2020): Artificial intelligence. Unleashing the power of AI for edu-cation. MIT Technology Review.