Artificial Intelligence in the Classroom

ARTIFICIAL INTELLIGENCE IN THE CLASSROOM

Wynne Bel l

Bedford College of Higher Educat i on, Bedford, UK .

ABSTRACT

Various teaching strategies have been employed in attempting to overcome t he difficulties experienced by students learning computer programming on courses held at the Bedford College of Hi gher Education, England .

The probl em remains unsolved; the ma in difficulty encountered lies in the development of the algorithm, not in the syntax or semantics of the language . The contribution gained by the use of flowcharts has been negligi ble ; t he major contribution came from a llowing the s tudents to work in groups .

The essay follows the ar gument that, accept i ng the premise that programming requires logical thinking, a solution to this problem may be forthcomi ng if it i s tackled by helping these students t o develop problem solving ski lls from an ear l y age ; thereby placing the onus on t he

schoolteacher .

Use of educational tool s , name l y BIGTRAK, initially for the very young, followed by the TURTLE with LOGO programming , and l atterly microPro log is advocat ed .

The paper is introduced by a br i ef discourse on the concept of knowledge, in order t o confirm that to teach thinking i s difficult, and hence that there is a role for some aspect of artificial intelligence in the classroom.

By definition , a t eacher 's role in the classroom must be that of one who teaches , i mparting knowledge and guiding the studies of the pupils. But, the concept of knowledge is difficult to define and cannot be determined precisely in the way that some words can. In attempt ing to define the constituents of the concept of knowledge, various notions can be considered, for exampl e : information, instruction, enl i ghtenment, learning , or practical skills.

Attempts have been made to break knowledge into divisions, Hirst,(l973) put forward the suggestion that knowledge was separable into distinct forms, such as mathematics, physical sciences etc. But, if these forms are accepted then the concept of a particular branch of knowledge may even differ, depending on how it i s presented . Polya , (1973) states, "Mathematics presented i n the Euclidean way appears as a systematic deductive science , but mathematics in the maki ng appears as an experimental inductive science ."

The view that the four ways of thought : logical, empirical, moral, and aesthetic represent more fundamental divisions of knowledge, Phi lips,(l971) perhaps relates more closely to the aim of including more of t he cognitive aspects of teaching in the classroom.

While the c l aim that knowledge i s worthwhile on its own account, s i mp l y for the deve l opment of the mind , can be appreci a t ed , as Gribbl e (1969) argues , forms of drill are not intrins i cally worthwhile . And while thi s form of t eaching perhaps cannot entirely be dispensed with , the question often arises relating t o whether teachers actually teach children to think, i.e . to deve lop their ability for reasoning, experimenting , making mora l reflection, or to achieve an appreciat i on of aesthet i c principles. The knowledge that i s planted in a child has to be brought into action .

Landa (1979) cons i ders that some teachers do teach children to think but that some do not. He admits that to teach thinking is a problem, because t he operations that have to be carri ed out on the knowledge present in the student's head , in order to be able to problem solve are not well deve loped .

The as sertion that organised teaching is not required for l earning t o take place, i s held by Papert (1980), who puts f on;ard the anal ogy of how

a child learns t o talk .

The difficulty of teaching children to think then is evident, and while the computer can be a catal ys ing agent for promoting a different type of teaching to take place in the classroom, to-date the methods employed in computer-aided-learning, have on the whole, contributed little t o further the pursuit t o help children to think . Computer-aided learning techniques employed i n schools generally follow the traditional method of classroom teaching , and if a chi l d 's development partially depends on this traditional approach t o teaching , then present computer-aided-learning does seem to provide mot i vation and have a novelty

value, albeit, poss i bl y a temporary innovation.

But,Papert's (1980) association of this use of a computer for drill and practice,(combined with the use of the BASIC language), with that of the Q\>IERTY keyboard , exemplifies the dangers i mminent when tradition takes a firm hold, and the ensuing difficulty encountered when endeavouring to bring about any change ; in this case in the · way that computers are used in the classroom.

Intelligent tutoring systems cater for the s tudent more adequately, and aid research into learning, but are rarely found in schools . Also artificial intelligence programs which deal with aspects of human behaviour designed to simulate behaviour help to illuminate how children think .

In the classroom, a simple machine in the form of a toy tank, BIGTRAK, can be used by primary and infant teachers for mathema tics teaching and to involve the children in logical thinking , besides providing an introduction to computer techniques . The tank can be programmed to move , turn, pause or fire and the children can think of their own problems and also how to solve them, and hence program the tank. They often act as BIGTRAK themselves in finding out the required movements, then logically assemble them to produce the program of instructions , which importantly, they realise may not be correct at the first trial. Hence , this toy enables a means of pupil- controlled investigation to take place , besides providing a by- product of introducing measurement and direction to them. The guidance of the teacher is required, as without this its value would diminish .

BIGTRAK then forms a medium for Papert ' s ideas,(l980) a lthough a rather unref i ned tool in some respects, learning is achieved through its use without formal teaching, and its advantage i n school s lies in its comparative cheapness.

The object that Papert(l980) advocates for use with chi l dren , the TURTLE , has a cost disadvantage at present, but with the infiltration of the microcomputer in all schoo l s in Britain , the concept of it, as an educational tool in the classroom, is rapidly gaining acceptance, even though many schools have to be content with " turtle" graphi cs on the screen , and it is expected that the cost will be lowered . But , even with " turtle" graphics, the child is in control and observations show that they usually enact the steps required to solve their particular problem.

The program l anguage incorporating the use of the TURTLE, LOGO, evolved for applications by children by Feurzeig et a l. (l969). The full version of LOGO provides additional facilities to the usual high l evel l anguage , for example, li st processing and recursive functions . But , in many schools in Britain onl y a subset of the l anguage is in use which relate only to controlling the screen "turtle". This reduces the amount of storage required and the cost of the software. Programs are also written in BASIC which can provide a reasonably successful LOGO environment in the classroom.

Hhile many school teachers in Britain are just beginning to be introduced to the potential of the use of LOGO as an educational tool , possibly because it was devised for such a purpose , a reasonable amount of research and evaluation has been carried out .

Preliminary observations in primary schools in Bedfordshire have revealed keen inte~est by both teachers and children in its use . The teachers have been impressed by the strategic skills shown by many of their pupils .

In their Evaluation Study : Teaching ~lathematics through LOGO Programming, (Howe et a l. 1980) , the conclusion reached is that the understanding of ma thematics by children who are less able can be improved by such programmi ng- based activitie s .

An investigation into the claims made for the use of the TURTLE in the classroom, involving 15 special schools situated over a wide area of England and Wales terminates in July 1983 . The research is co- ordinated by the Chiltern Advisory Unit, Hatfield , England; the final evaluation report should help teachers in deciding t he contribution that the TURTLE can make to the development of children .

While it has been shown that girls generally do not achieve as high results as boys in computer studies, the opinion is emerging that they do marginally better than the boys when LOGO is used .

But , it is not only chi l dren who can benefit from using LOGO, du Boulay (1978) showed that student teachers who exper i enced difficulty with certain areas in mathematics gained a better understanding by wr1t1ng LOGO programs to invest i gate the topics .

LOGO i s used in British and American schools and developments in France , to ascertai n it s potential for use as an intrinsic part of their educational system, are taking p l ace .

PROLOG-PROgramming in LOGic, designed by Colmerauer and colleagues in 1972 , i mages human reasoning and utilises natural language. Papert enthusiasts support the use of PROLOG for chi l dren ; a close relationship exists between LOGO and PROLOG.

PROLOG is now being made ava ilable for microcomputer systems in the form of microProlog, developed by Me Cabe 1980 . Relatively little work has been carri ed out on its use in t he classroom, but , a project has been running since October 1981, "Logic as a Computer Language for Children", based at Imperial College , London. Eva l uation is being conducted in a number of schoo l s and colleges , and courses are being held for teachers in various parts of Eng l and.

This project is led by Robert Kowalski considers that mi croProlog contributes promoting logical thinking for use throughout who to the school curriculum and that it can stand as a subject on its 01-m. lie considers that because it is not tied to a particular machine structure, it is more suitable for use by children than languages which are . (Ennels,l983) .

Enne ls (1983) has expressed surprise at the quickness that children are learning microProlog . The pupils build their own database and formulate queries , so promoting clear thinking and expresssion .

These then are some of the tools that can be used in assisting teachers to teach children to think ; many other micro-technology aids are available and although many are considered to be simply computer toys, an investigat ion into their potential use in the classroom may reveal that some are more than toys. "Computer toys come closer to imitating the style of human intelligence than the teaching machines of the past and may 1-1ell represent the educational 1-1ave of the future. " (Gardner,l979).

In conclusion, the educational tools suggested for use in this paper are mainl y just beinr, introduc.ed to schools in the Bedfordshire area, hence it will be some time before there can be any evidence to show that t he children 's capacity for clear thinking has improved and as a result the difficulties experienced by the ones who may eventually wish to include further computer programming as part of their future studies may be lessened. Current research in this area suggests

that this will be so .