Professional Documents
Culture Documents
Integrating Artificial Intelligence Into
Integrating Artificial Intelligence Into
To cite this article: Amelia A. Baldwin-Morgan (1995) Integrating artificial intelligence into
the accounting curriculum, Accounting Education: an international journal, 4:3, 217-229, DOI:
10.1080/09639289500000026
Taylor & Francis makes every effort to ensure the accuracy of all the information (the
“Content”) contained in the publications on our platform. However, Taylor & Francis,
our agents, and our licensors make no representations or warranties whatsoever as to
the accuracy, completeness, or suitability for any purpose of the Content. Any opinions
and views expressed in this publication are the opinions and views of the authors,
and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content
should not be relied upon and should be independently verified with primary sources
of information. Taylor and Francis shall not be liable for any losses, actions, claims,
proceedings, demands, costs, expenses, damages, and other liabilities whatsoever
or howsoever caused arising directly or indirectly in connection with, in relation to or
arising out of the use of the Content.
This article may be used for research, teaching, and private study purposes. Any
substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing,
systematic supply, or distribution in any form to anyone is expressly forbidden. Terms &
Conditions of access and use can be found at http://www.tandfonline.com/page/terms-
and-conditions
Accounting Education 4 (3), 217-229 (1995)
Abstract
The development and use of artificial intelligence technologies continues to expand in accounting
practice, industry and government. However, the exposure of accounting students to artificial
intelligence remains almost non-existent. Accounting faculty should move towards integrating
artificial intelligence topics into the curriculum so that students can begin learning about artificial
intelligence before they encounter it in the workplace. Artificial intelligence techniques give computer
systems human-like abilities to reason and learn. For accounting, the most promising types of
artificial intelligence are expert systems and neural networks. Artificial intelligence is used in
auditing, financial and management accounting, taxation and government. The integration of
artificial intelligence into the accounting curriculum befits the move for microcomputer integration
and calls for the innovative use of technology in the accounting courses. Suggestions from the
literature include using supplemental readings and requiring students to use and/or build simple
expert systems. The motivations and examples discussed here should provide the impetus for other
instructors to integrate coverage of artificial intelligence topics into their accounting courses.
Keywords: accounting education, IT integration, artificial intelligence, expert systems.
Introduction
Artscial intelligence is one of those seductive terms that brings to mind visions of a future
world where everyone has a life of leisure and only artificial lifeforms carry out the
drudgery of day-to-day labour. Well, that time is not yet upon us. However, artificial
intelligence is becoming a more common technology, with commercial applications in the
accounting domain. As this technology gains widespread use, accounting students must be
exposed to its features and potential before reaching the workplace.
Actually affecting student knowledge is usually the last phase in a seemingly straight-
forward process of curriculum change. Four steps can be followed to integrate artificial
intelligence into the curriculum (Figure 1). First, faculty must recognize the need for
artificial intelligence integration, that is, the need for future accountants to become
familiar with artificial intelligence. Second, faculty must become informed about artificial
intelligence. One of the biggest barriers to change is the lack of accounting educator
knowledge about artificial intelligence. Third, knowledgeable professors and departments
can then intelligently implement appropriate changes to the accounting curriculum.
* Address for correspondence: Amelia Baldwin-Morgan, Department of Accounting, 406 Owen, College of
Business, Eastern Michigan University, Ypsilanti, M1 48197, USA. Email: acc~abaldwin@emuvax.emich.edu
0963-9284 0 1995 Chapman & Hall
A. A. Baldwin-Morgan
Recognize
need for AI
\
.-U
- --
Downloaded by ["Queen's University Libraries, Kingston"] at 16:29 07 March 2015
knowledge 1 )
Fig. 1. Steps for integrating artificial intelligence into the accounting curriculum.
Fourth, the enlightened professors and the changes in the curriculum will affect student
knowledge about artificial intelligence.
The purpose of this paper is to address the three steps needed to integrate artificial
intelligence topics into the accounting curriculum and to assist faculty in this process. The
paper proceeds as follows. First, artificial intelligence is defined along with the principle
areas that affect accounting practice. This basic knowledge should be a part of the
repertoire of each accounting educator. Second, examples of the use of artificial intelli-
gence in accounting and industry are given. These not only illustrate the variety of artificial
intelligence applications in accounting, but also could provide interesting material for
beginning a class discussion. Third, motivators for integrating artificial intelligence into
the accounting curriculum are identified. Fourth, a plethora of examples and suggestions
Integrating artfical intelligence into the accounting curriculum 219
for integrating artificial intelligence into the curriculum are described. Some final observa-
tions conclude the discussion.
Artificial intelligence
The American Institute of Certified Public Accountants (AICPA) provides this practical
1 definition of artificial intelligence:
a set of techniques whose processes give computers the human-like abilities to see,
hear, speak, reason with imprecise or incomplete information, and learn (1987,
Downloaded by ["Queen's University Libraries, Kingston"] at 16:29 07 March 2015
P 1).
The artificial intelligence abilities of reasoning and learning have the most interesting
potential for accounting applications. Currently, the forms of artificial intelligence most
likely to impact accountants are expert systems and neural networks.
Expert systems
Expert systems are computer systems that imitate the thinking processes of human experts
in order to solve problems in narrow domains (AICPA, 1987). The purpose of developing
expert systems is to have expertise available even when the human expert is not.
RULE :
IF Transfer-of-Ownership=Yes
AND
Lessee~or~Lessor~type=Lessee
THEN Type-of-Lease = CAPITAL-LEASE
Rule-based knowledge representation is the most common and easiest knowledge repre-
sentation scheme to understand. l
l
Some expert systems represent knowledge using frames rather than rules. Frames associate l
an object or concept with a set of characteristics (e.g. values, facts, rules, etc.). Each
characteristic is stored in a slot. A frame is the group of slots associated with a specific
object. A frame is somewhat analogous to a record, in database or programming terms
(Harmon and Sawyer, 1990).
Case-based reasoning
Case-based reasoning uses past experiences in the way that humans use experiential
reasoning to solve complex problems. This type of expert system is most appropriate
when little evidence is available or information is incomplete. The case base is made of a
set of cases. Each case contains information about its situation, its solution, the results of
that solution and key attributes. The index library searches the case base and retrieves
cases similar to the problem currently being addressed. Similarity metrics are used to
measure how similar a past case is to the current problem. Learning can occur by adding
new cases and their solutions to the case base (Brown and O'Leary, 1994).
Case-based reasoning is particularly suited for application to accounting principles
derived from prevalent industry practices, auditing, compliance accounting and manage-
ment accounting. Commercial case-based systems have been developed for financial credit
evaluation, bankruptcy prediction, credit evaluation (Mott, 1993), real estate property
appraisal (Gonzalez and Laureano-Ortiz, 1992), bidding for manufactured products, and
fraud detection (Brown, 1994).
l
Neural networks
Neural networks are computer systems and software that are built based on the organiza-
tional principles of the human brain (Duggal and Popovich, 1993). The brain organizes its
neurons to compute much quicker than the fastest digital computers (Kinoshita, 1988).
Although the brain's neurons are actually slower processors, the brain somehow organizes
them to make faster decisions. Neural network developers want to take advantage of that
organizational power of the brain.
Where traditional expert systems are rule based, neural networks are example based.
The knowledge in a neural network is derived from the data not a human expert (Ether-
idge and Brooks, 1994). Neural networks learn by example. Neural networks perceive
patterns by examining large amounts of data about examples. The neural network
classifies new examples based on the patterns recognized in them (Duggal and Popovich,
1993).
The finance industry has created the most neural network systems for business related
applications. Avco Financial Services analyses loan applications with a neural network.
Integrating artijical intelligence into the accounting curriculum 22 1
Neural networks created by Adaptive Systems assess mortgage loan applications. Nestor
and Wyman's neural network analyses consumer loan applications (Stipp, 1988). Chase
Manhattan Bank uses a neural network to detect credit card fraud (Gulbo, 1990). Nikko
Securities, Frontier Financial and Citicorp improve their securities trading strategies by
using neural networks. Kodak and a number of direct-mail companies also use neural
networks (Etheridge and Brooks, 1994).
Neural networks are appropriate for tasks whose results may be based on examples and
whose rules are not formally organized or even known. Explanation should not be a
requirement of the task if neural networks are appropriately applied. Accounting tasks
that may be suitable for neural network development include audit error identification
Downloaded by ["Queen's University Libraries, Kingston"] at 16:29 07 March 2015
l
(Duggal and Popovich, 1993), bankruptcy prediction (Etheridge and Brooks, 1994), and
loan and credit evaluation (Duggal and Popovich, 1993). Accounting firms are investigat-
ing neural networks. An audit firm was reported to be developing a neural system for
identifying audit errors (Duggal and Popovich, 1993). KPMG Peat Marwick uses a neural
network to predict bankruptcy (Etheridge and Brooks, 1994).
Neural networks eliminate some of the problems associated with rule-based expert
systems. The expertise in a neural network need not be formally organized. Also, neural
networks are very adaptive, unlike rule-based expert systems. However, neural networks
do have limitations. For example, they do not explain how solutions are achieved; no
decision rules are evident (Duggal and Popovich, 1993).
expert systems are used for corporate, international and individual tax planning, tax-
exempt status maintenance, compliance checking, uniform capitalization, tax issues con-
cerning preferred shares, and resident alien status questions (Brown, 1991).
On the financial front, accounting firms use expert systems to provide personal financial
planning to individual clients (Phillips et al., 1990). Shearson Lehman's neural network
system predicts stock and bond performance (Schwartz and Treece, 1992). The State Bank
of New South Wales, Australia, uses an expert system to assess loan applications (Zawa, 1
1994). ~
Expert systems are also used in management and management accounting. Some
companies use expert systems to provide personal financial planning assistance to their
Downloaded by ["Queen's University Libraries, Kingston"] at 16:29 07 March 2015
employees (Phillips et al., 1990). American Express uses an expert system to authorize
customer purchases. IBM's expert system assists its field service representatives in pricing
bids. Texas Instruments has an expert system that helps employees with capital budget
requests (Leonard-Barton and Sviokla, 1988). Another neural network identifies Spiegel's
most likely buyers (Schwartz and Treece, 1992).
Artificial intelligence is also used by governmental entities. A California county govern-
ment uses an expert system to determine eligibility of applicants for welfare benefits. A
similar system was developed for the state of Pennsylvania (Schwartz and Treece, 1992).
Applications of artificial intelligence to accounting-related tasks are developed by a wide
variety of organizations. The systems just mentioned were developed by public accounting
and consulting firms, federal and local governments, commercial developers, and corpo-
rate developers in a variety of industries.
External motivators
The most important reason for including artificial intelligence in accounting education is
to better prepare the students for the professional world. According to Bandy (1990), that
is the purpose of accounting education, not to teach accounting to students, but rather to
teach students to become accountants. Hatherly and Fraser (1991) concur. They assert
that because expert systems are being used in the profession, accounting students should
become familiar with these tools.
Internal motivators
The integration of microcomputers in the accounting curriculum is required by the
American Association of Collegiate Schools of Business (AACSB), the principal accred-
iting body of business schools in the United States. Currently, about 25% of schools of
business have been accredited by the AACSB. Less than 15% of accounting programmes
have obtained separate accreditation from the AACSB. The use of expert systems in
accounting education is an innovative method for meeting the AACSB's microcomputer
integration requirement (1985).
Another inside motivator that is specific to accounting is the Accounting Education
Change Commission (AECC). The AECC, in its first position statement, encouraged the
creative use of technology. Further, the AECC asserts that accounting students should
learn by doing (1990). Using and building simple expert systems in accounting courses
Integrating artijical intelligence into the accounting curriculum 223
appropriateness of expert systems use for the training of accountants has been supported
by both practitioners and academicians (Black et al., 1990; Jancura, 1990; Englard et al.,
1989; Jancura and Overbey, 1988; Gal and Steinbart, 1987; McKee, 1986; Elliot and
Kielich, 1985).
General suggestions
White (1995) suggests seven ways to cover expert systems in class. The least difficult
method in terms of time and resources is the textbook approach. Most accounting
information systems textbooks have some discussion of artificial intelligence. Often, this
is a mention of expert systems in the context of a chapter or segment on decision support
systems. Some instructors may desire to augment the textbook coverage with supplemental
readings such as Brown and O'Leary (1994), which has been distributed to members of the
Artificial Intelligence/Expert Systems special interest section of the American Accounting
Association expressly for use by faculty and students, or journal articles such as Brown
and Murphy (1990) or Brown (1991).
With a bit more effort, accounting teachers may require a project using a sample expert
system. Those more industrious could ask the students to build a simple expert system
using a simple expert shell. Others may want students to build a complex expert system.
Only the most ambitious professors will have accounting students learn an artificial
intelligence program language or take a course in artificial intelligence programming
(White, 1995).
While White (1995) was envisaging coverage of artificial intelligence and expert systems
in the accounting information systems course, coverage in other courses has also been
suggested. For example, instructors in a number of accounting courses may use supple-
mental readings that are geared specifically to their area of interest. Most of the reading
suggestions listed below were published in practitioner journals or journals meant for both
a practitioner and an academic audience. Most are not lengthy, and many include a brief
section defining and describing expert systems technology. Those listed below represent a
selection of the more recent articles that may be appropriate for student consumption.
Tax instructors may use McDuffie et al. (1994), Schreiber (1992) and Black et al. (1990).
I Auditing students can read about expert systems use in internal and external auditing in
224 A. A. Baldwin-Morgan
Sriram (1992), Brown and Phillips (1991) and Brown and Murphy (1990). Management
accounting instructors may assign readings by Brown and Phillips (1990) and Brown and
Phillips (1994). Brown and McDuffie (1993) may be used in a governmental accounting
course. Several articles may be particularly appropriate for use in MBA-level courses, e.g.
Schwartz (1992), Yamasaki and Manoochehri (1991), and Leonard-Barton and Sviokla
(1988).
Examples of use in accounting courses
For instructors who want to go further than using supplemental readings, several choices
for integrating artificial intelligence and expert systems into their courses have been
Downloaded by ["Queen's University Libraries, Kingston"] at 16:29 07 March 2015
mentioned. The two most common involve having students use an expert system and
having students build an expert system. Use of an expert system by students will give them
a basic understanding of the technology and of the expert system's task. Building a simple
expert system carries the added bonus of providing a deeper understanding of both the
technology, the task requirements, and problem-solving skills. Students who build an
expert system also gain some understanding of the process of system development.
Expert systems have been introduced in almost every possible accounting course. Boer
and Livnat (1990) suggest that teaching accounting topics with an expert system is
appropriate when the accounting issue has a complex set of rules which can be divided
into segments and have an ordering in the decision process. Further, these routines must
need practice to be mastered fully. Examples of how instructors have used the technology
and additional methods to try are given in this section for specific portions of the
accounting curriculum.
Financial accounting
Boer and Livnat (1990) demonstrated the use of an expert system project to teach
intermediate accounting students how to classify capital or operating leases. One class
used the expert system (built with Exsys) and another class used a traditional textbook
approach. The expert system users made fewer errors. Boer and Livnat (1990) concluded
that expert systems tools can be an effective teaching aid.
Sangster and Wilson (1991a,b) described an experiment that involved students con-
structing and using an expert system applied to three different UK Statements of Standard
Accounting Practice (depreciation, research and development, and investment property).
Their conclusions definitely support the use of expert systems by students. First, having
students develop and use their own expert system tools required less preparation time than
non-expert-system computer-aided learning. Second, the development and use of the
expert systems accelerated student comprehension. Third, the expert systems project
required students to exercise their problem solving skills. Fourth, student exposure to
expert systems helped them appreciate the potential usefulness of expert systems in the real
world.
McDuffie (1990) constructed and validated an expert system (using VP-Expert) for
accounting for business combinations. This system applies the criteria identified in APB
Opinion No. 16 to the purchase or pooling decision (McDuffie et al., 1993). The developer
has shared his prototype expert system with others interested in expert systems in
accounting (Smith et al., 1991).
Booker and Kick (1987) developed an expert system for use in intermediate accounting
using Expert Ease. Their article describes the development process in detail. The system
Integrating artijical intelligence into the accounting curriculum 225
applies Accounting Principles Board Opinion No. 21, Interest on Receivables and Payables.
Many other expert system developments are possible in financial accounting. Hatherly
and Fraser (1988) discussed the development of an expert system for classifying inter-
company investments. Boer and Livnat (1990) suggest other financial accounting topics
that may be appropriate for expert systems use, such as earnings per share, accounting for
pensions, foreign currency translation, business combinations and accounting for deferred
compensation.
management accounting. The knowledge area encompassed in the system was standard
costing.
Taxation
French and Flesher (1987) suggest that tax educators must prepare for the impact of expert
systems on tax education. However, few have expert systems been reported in use in tax
courses.
Bouwman and Knox-Quinn (1995) had masters students in taxation build simple expert
systems. The students improved not only their knowledge about the tax subject, but also
their problem solving skills. Hye and Nar (1993) described the development and testing of
an expert system for teaching students in Singapore how to segregate tax deductible
expenditure items from non-deductible items.
Auditing
Auditing is one area that has seen much expert systems development in public accounting,
but no corresponding coverage in auditing courses. Two of the just described systems,
while used in an accounting information systems course, covered an auditing-related task,
the evaluation of internal control. These give some idea of the potential for expert systems
use in auditing courses. Another method for introducing such systems to students is to
have a demonstration of one by a representative of a public accounting firm. Whether
firms are willing to provide this service or to allow semi-public demonstrations of their
technology is not known.
Downloaded by ["Queen's University Libraries, Kingston"] at 16:29 07 March 2015
Conclusions
The increasing use of expert systems for accounting tasks in accounting practice, industry
and government is a signal that future accountants should begin learning about artificial
intelligence during their undergraduate studies. This paper presented information intended
to assist accounting educators in becoming knowledgeable about artificial intelligence, and
in understanding the need for incorporating such knowledge into the curriculum. In
addition, examples and suggestions from the literature were provided to assist in curricu-
lum change.
Changing the curriculum is never an easy task. Witness the ongoing effort to integrate
microcomputers into the curriculum. Even in the 1990s, some accounting instructors
remain unwilling to change (King and Whittaker, 1991). Like microcomputers, integration
of artificial intelligence topics into the accounting curriculum will not happen quickly.
However, one method for combating lack of knowledge and unwillingness to change is the
dissemination of useful information and the existence of successful exemplars. Hopefully,
the information contained herein will help to push change in the curriculum one more step
forward.
Acknowledgements
The author is grateful for the helpful comments of Gary McCombs, Mohsen Sharifi, and
Skip White, the anonymous reviewers and the guest editor.
References
Accounting Education Change Commission (1990) Objectives of education for accountants: position
statement number one, Issues in Accounting Education, 5 (2) Fall, 307-12.
American Assembly of Collegiate Schools of Business (1985) Accreditation Council Policies,
Procedures and Standards.
American Institute of Certified Public Accountants (1987) An introduction to artificial intelligence
and expert systems. A Management Advisory Services Special Report, (AICPA: New York).
Armitage, H.M. and Boritz, J.E. (1986) Integrating computers into the accounting curriculum, Issues
in Accounting Education, 1 (l), 86101.
Bagranoff, N.A. (1990) Using the computer in the accounting information systems course, Journal of
Accounting and Computers, 6 , Fall, 43-54.
Baldwin-Morgan, A.A. (1994) The impact of expert systems on auditing firms: evidence from a case
Integrating artzjkal intelligence into the accounting curriculum 227
Michaelsen, R.H. and Messier, W.F., Jr. (1987) Expert systems in taxation, The Journal of the
American Taxation Association, 8 (2), Spring, 7-21.
Mott, S. (1993) Case-based reasoning: market applications and fit with other technologies, Expert
Systems with Applications, 6 (l), 97-104.
Peek, G.S. (1991) Using expert systems projects in the accounting information systems course: the
rational and an example, Journal of Accounting and Computers, 7, Fall, 99-108.
Phillips, M.E., Brown, C.E. and Nielson, N.L. (1990) Personal financial planning with expert
systems: an expanding employee benefit, Management Accounting, 72 (3), 29-33.
Raval, V. (1989) A curriculum-wide approach to integration of computer in accounting education,
Journal of Information Systems, 3 (2), Spring, 132-44.
Sangster, A (1991) Expert systems in the accounting curriculum: a textbook review, British
Downloaded by ["Queen's University Libraries, Kingston"] at 16:29 07 March 2015