IEEE.

org IEEE Xplore IEEE SA IEEE Spectrum More Sites Subscribe Donate Cart Create Account Personal Sign In

Browse  My Settings  Help  Institutional Sign In

All  

ADVANCED SEARCH

Journals & Magazines > IEEE Access > Volume: 8 

The Impact of Artificial Intelligence and Blockchain on the More Like This
Accounting Profession Artificial Intelligence, Blockchain,
Publisher: IEEE Cite This  PDF Big Data Analytics, Machine
Learning and Data Mining in
Yingying Zhang ; Feng Xiong ; Yi Xie ; Xuan Fan ; Haifeng Gu All Authors Traditional CRM and Social CRM:
A Critical Review
89 47605 2022 IEEE/WIC/ACM International Joint
Cites in Full      Conference on Web Intelligence and
Intelligent Agent Technology (WI-IAT)
Papers Text Views
Published: 2022

 Open Access  Comment(s)

Research on Financing Mode of
Under a Creative Commons License
New Agricultural Operators Under
the Background of "Blockchain +
Abstract Abstract: Big Data"
Recent developments in technology have introduced dramatic changes to the practice of 2022 6th Annual International
Document Conference on Data Science and
the accounting profession. This paper provides a comprehensive review of current Business Analytics (ICDSBA)
Sections
developments in big data, machine learning, artificial intelligence, and blockchain utilized in Published: 2022
PDF
I. Introduction general business practice and by specialized practitioners in the accounting profession
worldwide. This paper explores the evolution of the accounting profession following these Show More
Help

II. An Introduction to recent technological developments and assesses the impact of future developments.
Big Data, Inherent challenges and opportunities posed by these new technologies pertaining to
Machine accounting professionals and accounting educators are also examined, including an
Learning, and increased demand for IT professionals with accounting experience as opposed to
Artificial accounting major graduates. Considering the dramatic changes and developments of AI
Intelligence applications in accounting, this paper reflects how all these technologies and the associated
Technologies and requirements of job candidates will affect the desired capabilities of accounting graduates
Their Applications and provides further discussion regarding what higher institutions and their accounting
In the Business graduates can do to adopt such changes.
(Accounting)
Field

III. Application of
New
Technologies by
the Big Four
Accounting
Firms

IV. Beyond Artificial

Intelligence – Accounting students are recommended to hone their programming skills.

Blockchain
Published in: IEEE Access ( Volume: 8)
V. Suggestions for
Accounting
Page(s): 110461 - 110477 DOI: 10.1109/ACCESS.2020.3000505
Professionals
Date of Publication: 08 June 2020  Publisher: IEEE
Show Full Outline 

Electronic ISSN: 2169-3536

Authors

Figures  Funding Agency:

References CCBY - IEEE is not the copyright holder of this material. Please follow the instructions via
https :// creativecommons . org / licenses / by / 4 . 0 / to obtain full-text articles and
Citations stipulations in the API documentation. 
Keywords SECTION I. 
Metrics Introduction
Footnotes
As the “language of business”, accounting was first established by Luca
Pacioli in 1494 [1]. Its main functions are to measure an organization’s
economic activities and communicate such information to related
stakeholders, such as corporate managers, creditors, consumers, and
regulators [2]. Recent developments in intelligent automation have
introduced dramatic changes to the practice of many traditional
professions, including accounting. According to a study performed by the
BBC, accountants rank 21 of a total of 366 occupations that are likely to be
st

eliminated due to the introduction of artificial intelligence (AI), with an

elimination rate of 96% [3].

Surveys of more than 120 internal auditors from the 2016 and 2017 KPMG
IT Internal Audit conferences reveal that nearly half of the representatives
confirmed the use of AI by their organizations, at least to a limited extent.
However, these surveys also demonstrate that 80% had no confidence in the
governance around the use of these technologies, and 70% admitted that
they did not know how their auditing methods would be affected [4].

The Big Four accounting firms have recently launched their own financial
robots that are capable of automatically recognizing data, entering invoices,
and generating financial reports. These financial robots are likely to replace
basic accounting clerks, allowing business managers with zero accounting
knowledge to make informed business decisions based on basic accounting
information [5]. Therefore, it is necessary to investigate the current
development of AI applications in the accounting profession. The increased
prevalence of AI within the accounting profession is likely to transform
current accounting practices and inform the education development of
future accountants. AI technologies are widely used in financial distress,
financial fraud, stock market forecasting, and auditing [6]. The
improvement of accounting technical and data analysis skills has received
attention from accounting firms [7] and is required in accounting education
courses [8], [9].

The main objective of this article is to explore, present, and discuss the
applications of AI and blockchain technology in the accounting field,
especially the opportunities and challenges of current applications of these
emerging technologies in the accounting field, particularly in the
development of and education for the accounting profession. The article
reviews five AI technologies, including natural language processing, machine
and deep learning, artificial neural networks (ANNs), social robotics
(human-computer interaction), and computer vision; eleven applications,
including Nuance, Cortana Alexa, AlphaSense, TensorFlow, Kensho,
Microsoft Cognitive Services, Skymind, IBM Watson, Accenture myWizard,
and Clarifai; fifteen financial scenarios, including voice recognition,
conversation AI services, phone fraud detection, real-time demand and
behaviour analysis, intelligent voice banking assistants, intelligent financial
search engines, stock index predictions, public sentiment information
management systems, statistical analysis, identification of the withdrawal
customer on ATMs, bank fraud detection, tax preparation, intelligent
automation strategy, pattern recognition in images, and retailer inventory
management; and the underlying framework of blockchain technology and
the construction of a blockchain auditing application platform.

The rest of this paper is presented as follows: Section 2 provides a brief

introduction of the technologies of big data, machine learning (ML), and AI,
as well as their applications in the business (accounting) fields; Section 3
provides a further discussion of the current use of AI by Big Four accounting
firms worldwide; Section 4 reviews the concept of blockchain and its current
application in the accounting field; Section 5 reflects how all these
technologies and the associated requirements of job candidates will affect
the desired capabilities of accounting graduates.

SECTION II.
An Introduction to Big Data, Machine
Learning, and Artificial Intelligence
Technologies and Their Applications
In the Business (Accounting) Field
A. Big Data
Big data incorporate four key characteristics: large volume, high speed, huge
variety, and uncertain veracity [10], [11]. The term “big data” comprises
both enormous amounts of data and the data analytic techniques
(algorithms) used to analyse these data [12], [13]. Following recent
improvements in data storage and analytic capacity, companies are now able
to extract business value from data to better understand their business
environment, consumers, and competitors. For better accounting, big data
are valuable as a source of financial data to support business decision-
making.

Big data demonstrably improve decision-making and prediction accuracy

[14]. For example, it is now possible to examine unstructured data such as e-
mail, telephone, and audit committee meeting materials [15]. Previous
literature identifies the use of the wisdom of crowds on social media
platforms to detect corporate fraud [16]. In addition, data visualization
software can transform massive amounts of data into information that can
enhance decision-making processes [17].

Following the widespread adoption of big data, data security has become a
serious issue prompting new regulations in many countries. For example,
the European Union (EU)’s “General Data Protection Regulations” (GDPR)
came into effect on May 25, 2018, specifying corporate requirements in
protecting users’ data, including business use and sharing. Within the EU,
the Payment Service Directive 2 (PSD2) addresses concerns about consumer
protection, the promotion of innovation, and the improvement of the
security of payment services. The UK Competition and Markets Authority is
currently implementing a program known as Open Banking, which enables
customers to share data securely with other banks and third parties via
application program interfaces (APIs) [18].

B. Machine Learning
Broadly speaking, ML is the science of computers running without being
explicitly programmed [19]. It applies a series of statistical techniques, such
as mathematical modelling, data visualization, and pattern recognition, to
conduct self-learning activities with input data to predict and understand
data trends and patterns [20], [21]. Recent applications of ML include
corporate revenue forecast analysis and investment decision-making. For
example, Two Sigma Investments LP, a New York City-based international
hedge fund, works with vast sets of big data from over a thousand diverse
sources and uses ML to build powerful investment predictive models [22].
Other applications of ML in the business world include the prediction of
consumers’ purchase intentions, as widely used by Amazon and Taobao
[23].

In the practice of management accounting, ML could assist in transaction

classification with the scope of the control function, such as in financial
planning and analysis (FP&A). The use of an application of ML technology
allows the prediction of transaction classification based on the analysis of
historical transactions. However, prediction quality relies on the quality and
inherent bias of the dataset being utilized [12]. An example of transaction
classification is an e-mail communication tool that classifies marketing and
promotion to consumers as “advertising expenditure” and employee
communication as “IT or communication cost”. ML technology can be
trained to recognize the difference and clarify each category with a pre-set
algorithm. In the practice of tax administration, the Guangdong Provincial
Taxation Bureau adopts ML approaches to identify suspected fraudulent tax
practices [24].

C. Artificial Intelligence and Supporting Technologies

AI is the outcome of successful applications of big data and ML technologies
to understand the past and predict the future based on enormous data [19].
Table 1 records selected applications of AI, followed by a further discussion
of AI applications.

TABLE 1 Selected Applications Based on AI Technologies

Table 1- Selected Applications Based on AI Technologies

The Nuance security suite filters every call and compares voice
characteristics, including tone, rhythm and accent, to a voice digital library
associated with bank fraud. The software can quickly flag suspicious calls
and alert agents in call centres to a possible fraudulent attempt [25].

Cortana is Microsoft’s personal intelligence assistant. It can connect with

users in real time to analyse their needs and behaviours in order to help
them manage assets efficiently and conveniently [26].

Alexa is Amazon’s personal intelligence assistant that can activate the user’s
workplace and application via voice. It also provides guests with easy access
to the service via voice. Amazon has teamed up with banks to give users
voice access to their bank accounts [27].

AlphaSense is an intelligent financial search engine that helps

professionals who need to cope with large amounts of data solve the problem
of information overload. It provides access to proprietary research databases
and includes semantic knowledge of the financial language and correlation
analysis to unlock valuable hidden information. AlphaSense collects millions
of documents, including regulatory filings, company profiles, news, press
releases, Wall Street research reports, and other uploaded files [28].

TensorFlow was developed and is maintained by Google Artificial

Intelligence Team Google Brain, a symbolic mathematical system based on
dataflow programming that is widely used in the programming
implementation of various ML algorithms. TensorFlow can be applied in
stock index prediction [29] and public sentiment information management
systems [30].

Kensho is a leading real-time statistical computing system and scalable

analytical architecture dedicated to providing market transparency to
financial institutions through advanced technology. Kensho leverages
massively parallel statistical computing, user-friendly visual interfaces, and
breakthrough technologies in predictive analytics to provide investment
professionals with the next generation of analytics platforms [31].

Microsoft Cognitive Services gives every developer access to AI without

the need for ML expertise. With just one API call, it can analyse the contents
of images, customize image recognition to meet business needs, and detect
and identify people and emotions in images [32].

Skymind uses AI technology to automatically adapt to new behaviour,

using untapped data sources to detect fraud cases. Skymind builds and
trains a neural network that uses historical transaction data to score new
transactions based on their likelihood of fraud. Human analysts are used to
review risky trades, and the results are fed back to the neural network to
improve detection efficiency [33].

IBM Watson uses 10,634 documents and was trained by 13 tax experts on
five different development models. After training, Watson suggests the
correct tax treatment three out of four times. Tax professionals use Watson
to improve customer service and identify deductions and deductions [34].

The Accenture myWizard intelligent automation platform embeds

advanced analysis, insight, and automation capabilities, and enterprises
combine business process-led application outsourcing services with problem
management, intelligent knowledge management, and system stability
routines. Enterprises reduce costs and increase productivity by
implementing a comprehensive intelligent automation strategy [35].

Clarifai is used in combination with a camera system to calculate the

number of items left on the shelf through item detection and identification
and thereby understand the selling speed of commodities. For example, if
there is too much of an item on the shelf compared to the remaining quantity
of the item in stock, it is time to eliminate it, and if it is out of stock, it is time
to increase the order quantity [36].

D. Robotic Process Automation

Robotic process automation (RPA) is a repetitive and automated process
developed from AI technology [37]. It can be used to imitate human
behaviour, send e-mails, complete spreadsheets, and record and re-enter
data for other tasks. It functions on the basis of prescribed procedures and is
incapable of discerning changes in conditions, such as the external
environment. By 2020, it is estimated that more than 40% of data science
tasks will be automated, resulting in increased productivity and broader
usage by citizen data scientists [11]. PwC’s 2017 RPA survey found that 30%
of respondents have at least begun incorporating RPA into their businesses
[38]. While the decision to adopt a joint or centralized operational model is
a function of organizational culture, Ernst & Young (EY) suggest six key
components required for most traditional RPA operational models,
including vision and standard setting, identification of RPA application
scenarios, performance value measurement with adjustment, raising
awareness and training, and integration across the whole platform [39].

A major application of RPA in the accounting field is related to taxes. For

example, following the creation of automation software, robots are
configured to perform repetitive processes, such as submitting applications
to a tax authority portal. Tax automation can free up corporate tax teams to
focus on higher-value work, such as research, planning, and analysis. Figure
1 illustrates several RPA applications between manufacturers, retailers, and
consumers [40]. For example, following the completion of transactions
between manufacturers and retailers, as well as between retailers and
consumers, the issue of invoice and account receivable reconciliation can be
facilitated by constant updates with inventory databases and bank account
systems. Certain countries, such as China, require invoice verification before
the confirmation of tax benefits. Depending on the invoice format, OCR
scanning technology can be used to “read” and “record” paper format
invoices and store such information in real time. In addition, RPA
technology can perform real-time invoice verification by constant contact
with the government tax office database. The feature of RPA technology aims
to provide constant verification across various databases, facilitating real-
time business communications.

FIGURE 1. - Illustration of RPA applications.


FIGURE 1.
Illustration of RPA applications.

Show All

In the financial accounting and document approval field, SMACC has

designed software called “AI Extractor” to extract financial information data
from invoices [41]. This software is capable of automating accounting and
control, payment preparation, three-way matching in procurement
processing, business intelligence, and data analysis. With the assistance of
ML technology, it offers superior automation compared to traditional
optimal character recognition (OCR) and robotics solutions [42].

While RPA follows rules-based practices that allow the software to aggregate
data, trigger responses, and initiate new actions, AI adopts new technologies
such as voice and facial recognition to perform judgement-based responses,
succeeding RPA’s rule-based engines. When AI is integrated with RPA, it
allows the automation process to begin much more quickly, creating an
automation continuum.

1) Radio Frequency Identification

Radio frequency identification (RFID) utilizes wireless communication
technology to identify information stored on objects via electromagnetic
field reading. Following recent technological advances in reading and storing
information, RFID has become increasingly affordable [43]. One application
of RFID is to manage inventory and curb inventory wastages, thereby
reducing overall losses [44]. Developments of RFID technology contribute
primarily to increased process automation, such as inventory receipts,
shipments, inventory, and valuation. This facilitates increased efficiency by
RPA in the preparation of transaction ledgers and financial statements.
RFID provides information- sharing support for buyers in the supply chain
to improve the accuracy of purchase prediction. RFID uses smart shelves to
improve the flexibility of operations and reduce the cost of inventory
management [45]. In 2014, Zara parent company Inditex decided to adopt
RFID technology to improve the company’s supply chain, eventually leading
Zara toward ultrahigh efficiency in which the process of design concept to
finished product could be completed in only 10 days; it also led to better
sales and profit [46].

2) Speech Recognition
Speech recognition technology converts dialogue content into computer-
readable input [47]. It contributes to the further expansion of AI
applications, such as consumer service inquiries and foreign language
translation. For example, Alexa, the virtual assistant developed by Amazon,
is capable of interacting via voice, streaming podcasts, playing back music,
making to-do lists, setting alarms, and providing real-time information, such
as news, weather, traffic, and sports information. Such speech recognition
technology allows AI administrative assistants to coordinate and arrange
meetings between team members and external parties, contributing to more
efficient accounting practices [48], [49].

For example, Maycur . com is the leading domestic enterprise travel and
expense management SaaS platform. It mainly adopts speech recognition,
automatically converts semantics into structured information through AI
training, and automatically completes the filling of the amount, time and
place of the fee type. Voice accounting is similar to a dedicated intelligent
reimbursement secretary. It only needs to speak in order to complete the
reimbursement. It can extract the key information from one’s voice, perform
automatic classification, and complete the reimbursement process easily
[50].

3) Natural Language Processing

Natural language processing (NLP) focuses on understanding unstructured
data (from human sources) as an application of AI. Examples of NLP include
text mining, manual text analysis, and readability analysis [51], [52]. NLP is
used to find evidence for strategy-making based on the market environment
and consumer activities.

Different from traditional internal auditing, NLP technology can

automatically process unstructured text information, systematically and
automatically retrieve and review the main points of review so that internal
auditors are free from heavy reading and review work. At the same time, the
language model can identify high-risk cases that do not meet the target
terms and issues and conduct preliminary screening for internal auditors so
that they can focus on high-risk cases and conduct in-depth tracking to
achieve efficient internal audit work [53].

Notes:

(1) Invoice Issue – Sales data gathered from database;

(2) Accounts Receivable Reconciliation and Receipt Verification – Access

bank information;

(3) Invoice Verification – Access government database; Invoice Processing –

OCR scanning;

(4) Expense Reimbursement – Verification of invoice with logical query

check;

(5) Inventory Measurement – Real-time, based on transaction records;

(6) Process Payment – Utility bills, etc.

4) Artificial Neural Networks

ANNs attempt to simulate human neuron networks so that computers can
learn things and make decisions in a human-like way. ANNs are created by
programming conventional computers as if they were interconnected brain
cells [54], [55].

The artificial brain function of the ANN tool is at a much more advanced
level than traditional computer linear logic, which is able to establish a
stable network connection weight between the business data input and the
accounting element record. Following the establishment of a case library
based on the learning of historical economic transactions, the accounting
information configuration of future economic matters can be automatically
realized and integrated into the enterprise reporting system [56].

SECTION III.
Application of New Technologies by
the Big Four Accounting Firms

A. Deloitte
Deloitte has developed several applications based on advanced technologies.
For example, Deloitte developed an insight-driven organization (IDO)
framework to help organizations achieve strategic goals. IDO embeds daily
analysis, data, and reasoning into the decision-making process, which
facilitates the scaling of projects across the organization to drive greater
business impact by translating growing data volumes into measurable
business value and creating long-term competitive advantage from existing
data assets. IDO can also help to improve the speed and quality of decision-
making while reducing decision cost, transforming clairvoyant-like decision-
making from a solely executive pursuit to one achievable by all employees
[57].

Based on deep learning technology, Deloitte has also developed a voice

analysis platform called the Behaviour and Emotion Analytics Tool (BEAT)
to monitor and analyse voice interactions. BEAT has three key functions.
First, it monitors the voice interactions of customers. Second, it is capable of
identifying high-risk interactions through NLP. The language model
algorithm is used to judge the preliminary data extracted from internal and
external information and then determine the regulatory compliance of the
signed contracts (see Table 2 for an illustration of this approach) [58].
Third, it can alert users to interactions that may have negative outcomes
(e.g., complaints or behavioural problems) and provide detailed information
about the reasons for their occurrence [56]. BEAT is able to analyse over 30
different languages and 30 different behavioural indicators and can be
customized to meet specific risk and user requirements [59].

TABLE 2 An Example of High-Risk Contract Identification by Deloitte’s NLP

Table 2- An Example of High-Risk Contract Identification by Deloitte’s NLP

Through NLP innovations, Deloitte has developed an automated document

review platform using cognitive technologies that can read and automatically
identify relevant information in a set of documents. The platform has broad
advantages that enable Deloitte teams to process all types of unstructured
information quickly and accurately [60]. Natural language generation
(NLG), as a type of NLP, is used by Deloitte to obtain computer-generated
text for tax purposes. The company processes upwards of 50,000 tax returns
annually for clients’ employees who have expat status or other complicated
financial situations. Using NLG, Deloitte provides in-house individual tax
services for over 50,000 employees and creates detailed narrative reports of
individual tax returns. Its tax professionals rely on these reports to provide
more targeted financial advice to clients during consultations [61].

B. Pricewaterhousecoopers
PricewaterhouseCoopers (PwC) has a diverse portfolio of industry-specific
and cross-industry data and analytics solutions. For example, PwC uses RPA
technology to collect data and determine the filing status of all entities,
review trial balance sheets, and convert data into tax bases. These
procedures facilitate the preparation and revision of tax returns, tax
payment submission, and responses to related parties’ enquiries [62].

For auditing purposes, PwC has its own AI audit lab to improve audit
quality, automation levels, and operational efficiency and maximize the
capability of AI technology to collect comprehensive information and data
for rapid and accurate analysis. In collaboration with H2O.ai, a company in
Silicon Valley, PwC has integrated AI technology into accounting practice
through the creation of the GL.ai robot. Building on the training algorithm,
GL.ai adopts ML technology to absorb PwC’s global knowledge and
experience to stimulate the thinking process and make decisions similarly to
an experienced auditor. GL.ai is programmed to check every uploaded
transaction in milliseconds and to identify anomalies and suspicious
transactions in the general ledger. It has therefore become an essential
competitive advantage of PwC and has increased the company’s corporate
value [63].

PwC has also been successful in leveraging NLG. By implementing Narrative

Science’s AI power engine “Quill”, PwC has been able to work with global
financial institutions to automate customer understanding (KYC) reporting,
a previously time- and resource-intensive task. Using Quill, PwC has been
able to help clients reduce reporting time by approximately 25%, saving
more than $1 million per year. In partnership with Narrative Science, PwC
has also developed automated narratives for anti-bribery and anti-
corruption (ABAC) reporting. ABAC reports traditionally require
experienced compliance professionals to mine vast amounts of data to
identify and record violations. PwC uses Quill to develop automated
narratives that highlight potential misconduct based on risk models and
data, reducing the time required to produce a report from a few hours to a
few minutes and improving overall report consistency and quality [64].

C. Ernst & Young

EY is committed to building a portfolio of industry solutions based on
injection and innovation and has publicly set six industry priorities: financial
services, life sciences, retail and consumer goods, health, electricity and
utilities, the public sector, and the government. For each of these target
industries, EY is currently developing a range of accelerators, assets, and
products supported by its big data platform [65].

In addition to RPA and NLP technologies, EY uses a series of new

technologies, such as drones, to satisfy new business requirements. In
addition to utilizing machine reading (such as QR codes and barcode labels),
EY also uses drones to assist with inventory observation and real-time
analysis (such as optical characteristic recognition). Drone data are
transmitted directly to EY Canvas, the EY Assurance global audit digital
platform that seamlessly connects more than 80,000 auditors [66].

EY also adopts NLP technology in several of its business operations. For

example, when the Internal Revenue Service (IRS) issues a new lease
regulation, instead of re-examining all pre-existing lease contracts, EY uses
NLP to extract information and a human-in-the-loop to validate the results
[67]. The AI system is three times more consistent and twice as efficient as
traditional human teams. To access potential purchasing synergies in
mergers and acquisitions, EY has also adopted an intelligent classification
engine to quickly compile relevant information, a significant improvement
from manually building spreadsheets and pivot tables and reading lines of
non-standardized accounts payable and accounts receivable data [68].

To improve professional efficiency, EY has further adopted ML technology to

detect fraud. EY’s Fraud Investigation and Dispute Service (FIDS) has
achieved an accuracy rate of 97% in identifying suspicious invoices using ML
technology [69].

D. Klynveld Peat Marwick Goerdeler

Klynveld Peat Marwick Goerdeler (KPMG) has implemented a strong data
and analysis vision within its core practices, with the Global Expert Insight
Center providing complementary resources to local teams. KPMG is
currently developing its data and analytics service portfolio within member
companies to help customers address specific data issues, such as privacy,
security, and forensics. It is also extending its current services to new
markets. KPMG combines its tax, advisory, and audit credentials with its
digital investments and uses its value delivery framework and trusted
information to help clients transform their businesses.

In 2017, KPMG launched a team called “KPMG Ignite” to focus on research

and exploration. It is a technical partner ecosystem that helps build and
provide solutions using AI with tested open source tools and accelerates the
development and delivery of AI solutions. To ensure the smooth operation of
KPMG’s AI technology, KPMG Ignite provides continuous testing for
prototype development and innovation, as well as a framework and guidance
to address employees’ and customers’ queries with AI applications. KPMG
Ignite functions as a data analysis processor that provides unbiased
estimations. To ensure its effectiveness and credibility, KPMG performs
regular inspection and maintenance. In addition, KPMG has established two
special departments to address Internet and data security. One department
is responsible for company software design and testing network security.
The other department addresses data protection and instant responses to
cyber attacks [70].

KPMG has also introduced a novel means of assessing risks. KPMG’s

dynamic risk assessment (DRA) combines actuarial theory, complex
algorithms, mathematics, and advanced data with analysis to identify, link,
and visualize four-dimensional risks (severity, likelihood,
interconnectedness, and velocity). Compared with traditional two-
dimensional risk assessment (high likelihood and severity), DRA takes into
account the risk interconnectedness and velocity with which the risk may
affect business operations.

Combining the latest applied science with management insights and

extensive benchmarking, DRA modelling allows audit professionals to see
the spread (“contagion”) of risks that could form key clusters or trigger other
risks. By identifying the expected contagion effect between global and
corporate risk, KPMG can objectively measure important threats. These new
insights provide a new level of risk assessment for audit professionals, help
improve audit quality, and provide the organization with information about
how best to address and monitor these threats [71].

Based on RPA technology, KPMG has developed “K-analyzer”, tax analytic

software that is capable of analysing thousands of transactions in a matter of
minutes. K-analyzer downloads data from corporate ERP systems to reduce
errors, uses automation to analyse a large amount of data, and then clearly
summarizes the results. This process generates a clear audit trail acceptable
by tax authorities and is capable of analysing tax-sensitive data in a cost-
efficient manner [72].

KPMG has also developed a robot to reduce the time required to complete
FBT compliance tasks, known as the FBT Automator. This tool is capable of
performing line project and ledger data analysis, including interactions with
business systems such as SAP and Oracle to access back-end data, and
preparing FBT for return to work files using numerical coding and fuzzy
word matching without manual line-by-line review. KPMG has also
introduced the Payroll Tax Automator tool, which automatically fills in the
payroll code and allocates wage codes to the correct the type of payroll tax
wage. It can also run data analysis across monthly payroll tax returns and
annual payroll tax adjustments to ensure compliance with the initial
submission and generate a central storage facility for payroll taxes that
functions as a data collection room [73].

The KPMG Automatic Exchange of Information (AEOI) reporting tool aims

to streamline the reporting process by using hundreds of data validation
checks to create and embed the required XML files for submission. KPMG
offers licensing solutions and hosting services to translate system data into
relevant reporting formats, such as the Foreign Account Tax Compliance Act
(FATCA) and Common Reporting Standard (CRS) [74].

Examinations of the use of BG, ML, and AI technologies by the Big Four
accounting firms reveal two common trends. First, the accounting profession
is increasingly investing in AI and its integration into core business; second,
the Big Four claim that AI is a key factor for future success in the accounting
field. The following section outlines predictions for future development in
these areas.

SECTION IV.
Beyond Artificial Intelligence –
Blockchain

A. Blockchain Technology
Blockchain can be described as a series of blocks used to establish or record
ownership of assets between parties [75]. Blockchain does not need an
arbiter [76] and can therefore facilitate direct trading in the private sector
[77]. It improves the efficiency of transactions that require multiple
validations and validations via more rapid checking between parties [78].
Table 3 records some of these examples.

TABLE 3 The Impact of Blockchain on Accounting and Auditing

Table 3- The Impact of Blockchain on Accounting and Auditing

It can be seen from the table that the impact of the blockchain on auditing is
reflected mainly in the following aspects:

a. Nontamperable features provide a reliable source of data to support the

auditing business while ensuring the unique sources of accounting data.
Once the data for the audited entity are entered into the blockchain
network, they are difficult to tamper with [79].

b. Distributed ledgers improve the authenticity and reliability of audit data,

significantly reducing the risk of data being attacked by the audited unit
[80]. Traditional audit data are stored on a centralized cloud server and
are highly vulnerable to hackers, resulting in file loss or data tampering.
The blockchain stores data in a distributed manner and uses multiple
nodes to back up data [81]. Even if a single node is attacked by a hacker,
it will not affect the consensus state of the data in the network as a whole
[82].

c. The timestamp feature increases the difficulty of tampering with the

data for the audited unit and provides a stable audit trail for the
development of the auditing business [80], [83].

False transactions and accounting fraud are the main sources of material
misstatement risk. The existence of the timestamp feature requires that
the data for the audited unit be revised to reach consensus with multiple
participants. This greatly increases the difficulty of providing fraudulent
financial data and improves the authenticity and reliability of the data
for the audited unit, thereby reducing the cost of false financial
information verification [84].

d. Network consensus has improved the authenticity, reliability and

timeliness of the data for the audited entity, laying a solid foundation for
continuous audit and real-time audit work [85]. On the one hand, it is
more convenient to obtain audit request information through the
blockchain network, and audit requests can achieve minute-level or even
second-level responses, which can save information collection and
finishing time, thereby improving audit work efficiency [88]. On the
other hand, the consensus mechanism for blockchain enables all data to
be jointly confirmed the first time, which can guarantee the timeliness
and accuracy of the data. Greatly improving the authenticity and
integrity of the data saves many inquiry and correspondence procedures,
thereby improving the efficiency of audit work and saving labour costs
[89].

e. The programmable feature enables auditors to write audit algorithms

and audit business processing rules according to different audit
application scenarios, laying a solid foundation for the automation of
audit work [90].

In summary, the impact of blockchain on auditing mainly includes the

following two aspects:

a. The non-tamperable, distributed ledger, time stamp and network

consensus features of blockchain will enhance the authenticity and
reliability of the data for the audited unit and reduce the verification cost
of the data for the audited unit [79]. For data for the audited entity in a
detrusted environment, auditors may use electronic credit procedures or
reduce procedures such as inquiries and letters, depending on the
circumstances, thereby reducing audit costs. At the same time, the
improvement of the authenticity and reliability of the data for the
audited unit may also reduce the need for third-party verification
information, such as auditing [92].

b. In the blockchain programmable environment, auditors can write audit

algorithms or audit business processing rules according to specific audit
scenarios and promote the intelligent implementation of audit work
[91].

Blockchain technology can simplify the transfer of any value (data, assets,
currency, and information) in real time in a secure and cost-effective
manner. It is extremely useful to conduct transactions that require multiple
validations and validations, contracts, and any type of record validation. This
technology can also make it easier to detect fraud and errors by providing
clear and transparent information about transactions, as no one can modify
records once they are uploaded. Blockchain can be described as a series of
blocks used to establish or record ownership of assets between parties.
Blockchain does not need an arbiter [76] and therefore facilitates direct
trading in the private sector.

B. Applications of Blockchain by the Big Four Accounting

Firms
1) Deloitte
Deloitte has Blockchain laboratories in New York, Dublin and Hong Kong,
comprising more than 800 professionals. These laboratories work with
international organizations seeking blockchain solutions, such as food
suppliers that wish to track products from farm to table, banks and
insurance companies seeking fraud detection, and car manufacturing and
leasing companies [93]. Deloitte has developed more than 30 blockchain-
related models for purposes such as digital identity, digital banking, cross-
border payments, trade finance, and loyalty and reward solutions, as well as
unique work in investment management and insurance [94].

2) PricewaterhouseCoopers
PwC has developed the Blockchain Validation Solution, which combines a
risk framework with proprietary continuous audit software. This tool is used
to discover long-term indicator patterns that are not obvious to humans,
both direct and predictive, and provide objective results. PwC keeps a log of
transactions with customers who are currently experimenting with
blockchain technology and applying controls and testing standards to its
applications, including major stock exchanges and digital wallet providers
that allow users to monitor, view, and test transactions in real time [95].

3) Ernst & Young

EY has developed Blockchain Analyzer, a blockchain-based auditing
technology that enables in-depth reviews of cryptocurrency transactions and
supports the auditing of companies that use cryptocurrencies. It is a
foundational tool for auditing blockchain assets, liabilities, equity, and smart
contracts [96]. Blockchain Analyzer is developed in conjunction with
Guardtime (an enterprise blockchain company) and Microsoft’s Azure
Blockchain team. The platform is used to create marine insurance contracts,
digitize trading rules and automate the entire process by using smart
contracts to reduce paperwork. Insurance companies can update their
information online and check the location of ships around the world anytime
and anywhere. The platform makes the collective database available to all
interested parties in real time by connecting their core capabilities [97].

4) Klynveld Peat Marwick Goerdeler

KPMG has partnered with Microsoft to create the KPMG and Microsoft
Blockchain Nodes, an innovative workspace that combines KPMG’s industry
and blockchain application knowledge with Microsoft’s technical expertise.
Blockchain nodes are used to build a close relationship with start-ups and
developer communities, demonstrating the utility of blockchain technology
for a range of businesses, such as healthcare and the public sector [98].
Furthermore, the KPMG Digital Ledger Services is a tool that helps financial
services companies apply blockchain technology, streamline automated
back-office operations, deliver faster and more secure transactions, and
reduce costs [99].

C. The Impacts and Future Development of Blockchain

for Auditing
When all transaction details are stored on blockchain, it becomes impossible
to change information because all stakeholders have the same information
and all changes are recorded (with real-time updates). Therefore, blockchain
naturally eliminates any abnormal records (which no one dares to do) and
makes reliable real-time auditing possible. It is likely to significantly
improve audit quality and efficiency.

The integration of blockchain and cloud technologies allows auditors to

download information on business adversaries in the private chain and
directly use the opponent’s information to confirm the data during the
reconciliation process. The application of shared ledgers in blockchain frees
auditors from repetitive transaction checking but focuses on complex
transactions and internal controls, confirms the validity of the digital
performance of physical assets, ensures that the contract is written in
accordance with accounting standards, and changes the scope and methods
of audit opinions [100].

Future research can consider establishing an auditing process based on

blockchain technology. For example, a continuously working audit robot is
able to complete the audit record layer, such as fetching raw data, issuing
audit test result reports, and data packaging and chaining; improve the
traceability of each block structure on the audit analysis layer for Interim
audit testing; establish an early warning mechanism; and issue an audit
inquiry list and audit abnormal item list, etc.

SECTION V.
Suggestions for Accounting
Professionals
Based on the above review of the current developments of AI, and
blockchain technologies in the accounting profession, focusing on the Big
Four accounting firms, we provide suggestions regarding future
developments for accounting professionals and their organizations.

With the rapid technological advances of the last few decades, accounting
professionals are currently required to have programming skills and to be
proficient in data analysis. They must be capable of understanding emergent
tools, interfacing with reporting techniques, and interpreting reports to
answer questions from authorities [101]. A study prepared by the
Association of Chartered Certified Accountants in the United Kingdom
regarding the future of the accounting profession denotes the key tasks of
accountants and their capabilities in this modern era 1. For auditors, since
RPA technology makes real-time and continuous audits possible, audit
professionals should be flexible and adaptable regarding changing audit
procedures and should be able to process real-time risk information [102].
For risk management practice, when AI accounting software is utilized to
conduct a large proportion of accounting work and make decisions
accordingly, professionals need to ensure that the algorithm design is correct
(not intentionally manipulated or incorrectly self-developed), particularly
with regard to fraud detection. This process requires consistent monitoring
by experienced accounting professionals. As such, the introduction of
advanced technology into the accounting profession is likely to reduce the
number of employment opportunities for accountants without programming
and analysis skills. It may be necessary (and desirable) for companies to
provide adequate re-training for existing staff.

Accounting firms and departments may adopt several approaches in

response to rapid technological developments in this area. First, companies
can establish the vision of technology-driven professional practice as an
overall organizational strategy, such as setting up separated centres of
excellence, specific coordination with business units, or different focus units.
Second, organizations may conduct a comprehensive assessment of current
protocols employed in the data development and collection process, conduct
a candid review of its automation and analytical capability maturity, and
develop plans to implement appropriate IT support processes. Third,
organizations should determine the value that the proposed technology may
create and review it periodically. This process includes developing formal
key performance indicators so that efforts are focused in the right direction
to produce tangible results. Finally, to ensure data accuracy and security,
companies can establish a governance board to define and practice
enterprise-wide data governance standards. This governance board can also
address the effects of potential business disruptions during cohesive
technological changes within the corporation.

There is no doubt that financial robots will replace humans in performing

basic accounting tasks in the future, as they are already an important
component of the accounting landscape. As a financial practitioner, it is
important to continuously improve one’s professional knowledge and skills,
including computer expertise, to complete more challenging tasks. At the
same time, it is necessary for education systems to respond accordingly by
incorporating a higher level of technology proficiency.

SECTION VI.
A Reflection on Accounting Education
Transformation (in China)

Recent media and reports from professional bodies and firms indicate the
high probability of job loss within the accounting and finance profession due
to developments in big data, ML, and AI [103]– [105]. Such observations
present both challenges and opportunities for accounting education. On the
one hand, it renews a longstanding debate regarding the goal of university
education, namely, whether it should comprise liberal education or
professional training. On the other hand, these technological developments
can free educators from repetitive exam marking and tutorial consultations
so that they can spend more time with their students. As an illustrative
example, the current state of accounting education in China is examined to
suggest future changes.

In 2017, more than 270,000 students graduated with an accounting degree

(bachelor’s level and above) in China [106]. For a high school graduate who
is interested in pursuing a bachelor’s degree in accounting in China, there
are two options: attending either a government-funded university or a
government-recognized education institute (managed by a local or
international university). The most common approach is to enter a
government-funded university through the Chinese college entrance exam
(“Gaokao”). For accounting and finance education, there are two main types
of government-funded universities in China. First, comprehensive
universities such as Xiamen University, teach both accounting and finance,
as well as arts, sciences, and engineering; second, universities such as the
Southeast University of Finance and Economics focus on accounting,
finance, and economic 2. Students in comprehensive universities have easier
access to liberal education due to accessibility to courses from other majors
(such as arts and computer science). However, students in universities
focused on accounting, finance, and economics are more likely to be able to
access professional training related to their majors due to the large number
of student enrolments in these majors. It therefore makes more sense for
educators to spend time designing practical content such as ERP
demonstrations for large groups of students, while small classes are ideal for
the in-depth discussion of case study materials.

With the increase in AI applications in the accounting profession, the

pressure for students to look for a job after graduation, and the balance of
focus between teaching and research for faculty members, it is crucial to
determine whether university education should be aimed towards liberal
education or professional training. Although students at top universities
achieve high marks in “Gaokao”, this does not necessarily mean that they
want to get a job in Big Four accounting firms or investment banks
immediately after they graduate. Do employers prefer graduates with a
liberal education?

Technology developments such as RPA and OCR have significantly reduced

the workload of repetitive tasks for financial accountants. These tools can
scan and enter invoices automatically, enabling real-time audits.
Accountants are now expected to perform more value-added tasks such as
financial management and data-driven decision-making. The future of (flat)
organizational structures requires accountants to understand corporate
operations and contribute to corporate governance based on their
professional accounting knowledge. In our opinion, if we believe that
accounting is a career more than a job, higher education institutions should
continue the practice of liberal education rather than professional training.
Although rapid technological developments have pushed accounting firms to
restructure their professional knowledge and to reform their service
approach, certain professional characteristics of accountants are unlikely to
change, such as ethical and emotional intelligence [107].

Recent developments in AI can be viewed as an opportunity for educators to

move one step closer to educational fairness. For example, the integration of
big data and ML technologies creates an education system that can input a
variety of tutorial explanation approaches for the same question that is
accessible to students 24/7. Such a smart education system can facilitate a
range of learning approaches, benefiting teachers and students alike.

According to Han [108], an ancient educator of the Tang Dynasty in China, a

teacher should “discuss ethics, educate profession, and clarify confusion for
students”. The adoption of a smart system made possible by big data, ML,
and AI to provide highly efficient tutorial consultation for students allows
faculty members to spend more time in discussion with students, such as
sharing their professional experience. More importantly, educators can use
this extra time to design teaching cases that teach students to use their
accounting knowledge to solve practical issues of financial management,
corporate governance, and ethical dilemmas.

Recent technological developments have necessitated the transformation of

accounting education not only in China but also around the globe. However,
it is important to reiterate that the possession of adequate characteristics is
essential for graduates to adapt to current and future technology changes
and even create them. Therefore, educators should firmly support liberal
education while also encouraging and providing opportunities for students
to experience these novel technologies.

SECTION VII.
Conclusion

The deep integration of emerging technologies such as big data, ML, AI, and
blockchain in the accounting field has introduced tremendous changes to the
accounting profession, such as reengineering accounting procedures,
reducing accounting information errors and distortions, improving
accounting efficiency, and promoting the transformation of accounting
career structures.

Because this article reviews new developments in accounting, future

applications of new technologies such as blockchain are also discussed in
this study. For example, with the help of blockchain audit application
platforms, large-scale and real-time automated audits can derived. At the
same time, considering the increasing adoption of technologies in
accounting, how to protect data privacy becomes an important issue. While
corporations can enhance their data security structures, regulators are also
expected to strengthen and implement regulations for associated crime.

The article presents extensive and profound integration between AI

technology, blockchain and accounting. While the top accounting firms are
embracing these new technologies and challenges, accounting professionals
(practitioners, educators and students) are expected to expand their
technological knowledge, creating more efficient accounting practices.

Appendix A
See Table 4.

TABLE 4 Summary of AI Applications by the Big Four Accounting Firms

Table 4- Summary of AI Applications by the Big Four Accounting Firms

Appendix B

See Table 5.

TABLE 5 Table of Abbreviations

Table 5- Table of Abbreviations

Authors 
Figures 
References 
Citations 
Keywords 
Metrics 
Footnotes 

IEEE Personal Account Purchase Details Profile Information Need Help? Follow

CHANGE USERNAME/PASSWORD PAYMENT OPTIONS COMMUNICATIONS PREFERENCES US & CANADA: +1 800 678 4333    
VIEW PURCHASED DOCUMENTS PROFESSION AND EDUCATION WORLDWIDE: +1 732 981 0060

TECHNICAL INTERESTS CONTACT & SUPPORT

About IEEE Xplore | Contact Us | Help | Accessibility | Terms of Use | Nondiscrimination Policy | IEEE Ethics Reporting  | Sitemap | IEEE Privacy Policy
A not-for-profit organization, IEEE is the world's largest technical professional organization dedicated to advancing technology for the benefit of humanity.

© Copyright 2024 IEEE - All rights reserved.