Impact of Artificial Intelligence on Economics, Finance and Accounting
“The strong artificial intelligence (AI) revolution is bringing significant changes in the world of economics, finance and significantly altering the accounting profession. From the perspective of accounting professionals a wise move will be to espouse the technological challenges and adapt to the new business and management requirements by developing new AI skill sets and competencies. The advent of artificial intelligence is a highly important event in the history of economics, finance and accounting and objective should be set to effectively harness the power of AI to enhance the economic conditions in which we all can flourish. Read on…”
Artificial intelligence (AI) is the imitation of intelligent human behavior and thought processes in performing of various tasks. The mindboggling processing of almost infinite data at super speed, enormous storing capability and the development of mobile devices using technology are transforming the global economic system fundamentally as steam power did during first industrial revolution with almost godlike capabilities. With the present growth of computational capabilities AI will soon have the skills and intelligence of the human brain. With the developments in machine learning (ML) and deep learning (DL), machines are performing human tasks right from calculations to car driving, speech recognition, legal services, radiology, vaccine research and used in factories and battlefields.
Apart from automating routine low-skilled jobs, AI is increasingly automating creative work performed by high-skilled workers. AI has potential not only to compliment human labour but to replace human labour in an entirety. AI is moving toward super intelligence, an intellect much smarter than the best human brains in practically every field, including scientific creativity, general wisdom and social skills.
“AI is moving toward super intelligence, an intellect much smarter than the best human brains in practically every field, including scientific creativity, general wisdom and social skills.”
The big data science, machine and deep learning algorithms such as Natural language processing (NLP), Support vector machine (SVM), genetic algorithms, clustering/classification-Nearest Neighbor (KNN) and Reinforcement learning (neural network, Bayesian network, Naïve Bayes), Random Forests (RF), Logistic Regression (LR), Naive Bayes (NB) and Convolutional Neural Networks (CNN) are enhancing AI capabilities which can be used in production of goods and services in all the industries and even changing the organization structures of firms using this technology. AI is helping to solve complex problems by creating new ideas and scaling creative efforts. AI’s great ability of self-improvement can lead to “singularities” where boundless computer intelligence and economic prosperity materialize in finite time. AI is the most profound technological advance in human history in the age of “4th Industrial Revolution”.
At the heart of AI revolution is data which is called as new oil. For deploying a machine learning system in organization, an infrastructure of data, collecting and mining the useful data is required. The construction of data pipeline is laborious and costly due to idiosyncratic legacy systems of businesses difficult to interconnect. In data management process, a raw data needs to collected, organize and analyze into information understood by humans like a text document. After organizing the data, it is stored in a data warehouse further used by AI based system to interpret and visualize the data.
AI is used in financial accounting, auditing, cost accounting, taxation and financial planning. The expert systems (ESs) used in the accounting discipline replicates human expert’s behaviour and expertise and transforms it into system of rules to accomplish accounting jobs and resolve accounting issues and facilitates accounting-based decision processes.
This article analyze impact of AI on economic theories and the future prospects of finance and accounting occupations, to explore the required new skill set and competencies, the way humans and machines could work efficiently and effectively together. Based on review of published literature and recently performed surveys and reports, it is an attempt to shed light on some possible development trends in an AI in the context of economics, finance and accounting. It is based on interviews of 50 senior professionals in finance and accounting to study the key characteristics of finance and accounting profession based on standardized and measurable set of variables based on the up-to-date descriptions of the tasks and required set of the knowledge and skills to sustain and grow in future AI environment.
Impact of AI on Economic Theories
AI is transforming the global industries in unexampled ways and reformulating walks of all fields of humane strive. Economics which was antagonist towards AI has started realizing the value of AI in solving humanity problems. The economists usually adhere to theoretical models despite having no value in practical world due to inability to predict unpredictable human behavior in response to rapid technological changes like AI and big data due to inherent prejudices and biases and take decisions based on inadequate knowledge. The economic theories of rational expectations and efficient market are based on assumption that people react rationally in the given situation. But crises like euro debt crisis and 2008 economic slowdown proved it wrong again and again.
An economic model does not work all the times similar to theories in pure sciences due to changing perceptions and mood swings of economic actors and falsification of assumptions based on past experiences. The implementation of AI in economics can change this situation for better. AI can be used to correct these issues by doing data driven modeling of human behaviour by doing sentiment analysis which can improve the accuracy of forecasting future bubbles in economy. By implementing AI techniques and statistics in behavioural economics, economists can predict human behavior under changing situations.
“An economic model does not work all the times similar to theories in pure sciences due to changing perceptions and mood swings of economic actors and falsification of assumptions based on past experiences. The implementation of AI in economics can change this situation for better.”
The central banks and policy makers by estimating correctly when a recession may irrupt can swiftly implement monetary and fiscal policies to mitigate the effects of business cycles. The changes in supply and demand can be predicted correctly to enact required changes in order to avoid economic downturns by harnessing real time data coming from social media mimicking the price mechanism and consumer sentiments. AI will help the government authorities to base their decisions on correct real time data. AI can significantly contribute in policy formulation, analysis and evaluation using real time data and can accurately predict human responses to policies. With these possibilities, AI will help economics to be closer to pure science.
Interaction between Artificial Intelligence and Economic Theories
| Economic Theory / Concept | Interaction with Artificial Intelligence |
|---|---|
| 1. Game Theory | The model of “self-programmed machine learning” is an entrancing model for the game theory. So far such models are applied for full information games. The advanced neural networks can compete and cooperate with other players and the learned behavior will create the equilibrium for game theoretic models. The prevailing tax policy models are restricted in accounting for individual behaviour responding to a change in tax policy. With application of reinforcement learning based AI algorithms, the actions of the computational economic agent can be deciphered from an abridged game economy where AI agents can maneuver the game mechanics against its creator’s expectations. |
| 2. Decreasing Marginal Returns | The AI data also experience diminishing returns to scale similar to other factors of production. The accuracy of the machine learning based model revamp according to rising training dataset but at diminishing rate. |
| 3. Make or Buy Decision | Every organization worried about amount of value creation using AI technology and whether to develop AI capabilities like data mining, labeling and model buildings and data analysis within the firm or outsource it to cloud vendors which resembles classic make or by dilemma. |
| 4. Standardization vs. Differentiation | The AI services vendors contest fiercely to provide standard products and cloud services along with some sort of differentiation like better speed and effective performance to eliminate the competition. |
| 5. Minimum Efficient Scale | The relationship between fixed costs and variable costs of ML products will increase or decrease minimum efficient scale. The firm needs to occur high fixed cost to develop its own customized solution instead of buying standardized products from vendors. The innovative pricing mechanism like auctions and differentiated pricing can be used by vendors to optimize its cost. |
| 6. Returns to Scale | The supply side returns to scale requires incurring a large fixed cost for designing AI system and a small variable cost for distribution. The demand side return to scale depends upon economies of scale, pricing of product and number of incremental customers and quality of the service post sell. |
| 7. Bounded Rationality | As per Economics Nobel Laureate Herbert Simon, human lacks perfect information and brain to process facts and take rational decisions which keep on changing as per mood swings. This is called as bounded rationality. Using AI person can collect information and use it for effective decision making. Due to super increment in processing power of computers as per Moore’s law the human bounds can become flexible. |
| 8. Efficient Market Hypothesis | As per efficient market hypothesis of Nobel Laureate Eugene Fama states that it is challenging for individual trader with incomplete and imperfect information to beat the efficient market. But with the advent of AI powered computer traders in the markets, the more efficient the market become and degree at which market is efficient depends on the amount of traders in the markets use AI based algo trading. |
| 9. Prospect Theory | As per prospect theory of Nobel Laureate Daniel Kahneman and Amos Tversky (1979) when human make decisions with known probability of result, they compare probable losses against profits to make decisions. But in case of AI based technology such decisions are taken by programmed machines challenging the applicability of Prospect Theory. |
| 10. Taxation Policy | In Harvard-Salesforce researcher’s game theory model, AI agents with varied skills and specializations assemble resources and then earn money by constructing real estate assets or selling the products and services among themselves which creates inequality in their incomes. An optimal outcome is programmed as a judicial mix of money and leisure time and then the actors make their own choices. The reinforcement learning models creates activities in humans like tax avoidance strategies mimicking real economic actors. This model also creates an agent imitating policy regulators who change marginal tax rates to obtain efficiency and equality in income. |
| 11. Labour Market | A data in a firm is useless in terms of decision making if not analyzed efficiently internally by some expert. So it is crucial to hire people with requisite skills and competence to deal with data to gain competitive advantage in the new age of AI and big data. The shortage of such skilled workforce is acute problem firms are facing creating heterogeneity in productivity as this new data science competencies percolates through the labour markets. |
Impact of AI in Finance
- Natural Language Processing (NLP) for Credit Appraisal & Document Intelligence: Natural Language Processing (NLP) allows banks and FIs to appraise the risk of a credit applicant, observe consumer sentiments about their brand and customer service across the internet, develop credit score for under-banked clients and superfast document search for business intelligence.
- Credit Scoring via Digital & Social Media Footprints: The financial institutes firms in developing countries like India can offer mortgages and monetary services using AI to flourishing middle-class natives with lessoned or no credit history by understanding risk involved by analysing their social media footprints using natural language processing (NLP) algorithm based software. The digital footprint of such customer’s such as usage of social media, data about their net browsing, geolocation information is quantified into credit scores using AI.
- Personalized Investment Advisory: The customer’s spending and investments patterns are observed through their financial transactions which are correlated with market trends evolving through market research through social media, news sites using sentiment analysis software to offer personalized investment suggestion to customers.
- Dashboard Collation & Customer Profiling: The NLP software select most suitable data as per customer’s profile and his financial needs and collate it with other data such as name or age and present it on the FI’s dashboard.
- Sentiment-Driven Customer Care: The sentiment analysis software enables the customer care executive to deal with the customer more effectively. A customer care group can better service the customers by arranging tickets with the specific sentiments of the customers.
- Risk Quantification & Scam Detection: AI transforms issues in terms of risk management, scam detection into as set of numbers which can be easily accessed by the bank staff to make proper decisions and actions.
- Brand Advocacy & Influencer Identification: By using AI based NLP software influencers who can advocate brand of institute in their community can be spotted through their social media footprints such as blogs. The analyzed trend data about investments, trading can be passed on to customers through such influencers.
- Enterprise Documentation & Mortgage Automation: The enterprise wide documentation pertaining to loans and mortgages can be automated using NLP and further integrated into existing system without affecting ongoing operations. The historical data can be used as training set and useful information can be extracted out of tons of documents. The output data can be incorporated in dashboards which can be speedily accessed by loan or mortgage officers. The AI system provides a dashboard where employees can access a loan or mortgage application simultaneously.
- Conversational Chatbots & Proprietary Market Search Engines: A Chatbot interface for the bank allows the bank’s customers to scout the required information on bank’s webpage and get answers to basic transactional questions. Banks can develop their own search engines for extracting useful market information from tons of public company filings, equity research reports for treading in heterogeneous financial markets based on their own investment strategies.
Case of Axis Bank’s Multilingual Voice Bot
During Covid-19 the operations in Axis bank were challenged during the initial phase of lockdown due to extreme spikes in volumes, increment in complaints and escalations transpired by customer anxiety, which was a repercussion of limited services to customers during the lockdown phase. This affected Axis Bank’s customer experience index and the cross-functional service levels. Axis Bank partnered with an AI-based SaaS voice automation platform Vernacular.ai to optimise voice AI solutions and automate customer interactions via an intelligent human-like dialogue. The company created a multilingual voice bot for Axis Bank that can converse in English, Hindi as well as in mixed dialect. The bot employs automatic speech recognition and natural language conversion technology backed by AI based algorithms to provide an enhanced customer experience through automation of the contact operations and can deal with tons of customer issues raised every day with higher scalability. The state-of-the-art, deep neural networks trained on thousands of hours of acoustic data for text to speech recognition were used.
The system was launched in July 2020. AI Voice BOT has catered more than 2.23 million customers at an industry best success rate of an average of 85% and above. Approximately 65%-70% of customers were being touched by the BOT and attempted to connect with the contact center. In case BOT cannot resolve customer issue, the call is delegated to the human experts minimizing the navigation time. The AI system and human expert work together to deliver a rich and satisfying experience to customers. In the next phase, the bank is working on expanding the language capabilities to 10 Indian languages and adds 27 more self-service options.
Impact of AI on Accounting
Post world financial crisis of 2008, the businesses are delving means of new opportunities for augmenting raising revenue and return on capital through added income, reduction in cost and innovative sources of value creation to become more competitive and sustainable.
1. Opportunities to accounting profession due to AI
AI is transforming the future of many professions and accounting is one of the most significant such occupation. Various new job opportunities and roles are emerging as professional competence is augmented by applying AI tools along with creation of new job functions for accounting professionals in organizational setup. The advocator of AI upholds this AI revolution as a leap forward to espouse future challenges while the antagonists look at it as a balk due to skeptical attitude of some accountants to adapt to new changes in the business environment.
The accounting and financial activities which are less susceptible to automation such as interfacing with stakeholders, managing and developing people, applying expertise to decision making, planning, and creative tasks. The accountants can gain by using their AI competencies to solve broad issues, support decision-making by providing better and cheaper data, doing rigorous data analysis, imparting new insights on business and focusing on more valuable tasks after freeing up from routine working due to AI applications. There is a scant risk of job displacement for accounting professionals if they adapt to new technology, if accountants adopt new skills according to their changing role in the organization.
2. Challenges for accounting profession due to AI
- Frey and Osborne examined effects of AI on 700 plus by categorizing them as high, medium and low risk professions setting probabilities thresholds at 0.8 and 0.2. In this study the accounting profession is classified as risky in terms of job displacement due to AI with probability of 0.92.
- AI systems will substitute accountants in day to day routine activities by doing it speedily and precisely than human beings posing a threat in terms of job displacement.
- A practical challenge in converting rules and regulations into if-then rules and decision trees to be exercised in AI logic.
- The diverse and complex nature and large volume of AI and Big data makes it imperative for accountants to acquire new job skills for big data analytics.
- Technical practicability, higher outlay for setting automation, structural changes in labour markets, regulative and acculturational acceptance are essential elements luring speed of automation in accounting and finance.
- In changing work culture the productivity from combined efforts of human labour and machines will depend on organization’s structure and culture, patterns of business models and competition in idiosyncratic industry.
- The key challenges for policy makers are encouraging investment in new technologies and framing policies to assist labour force and organisations to deal with adverse impact of adapting AI automation if any in future.
“The leadership skills like strategic thinking, coaching, mentorship, morality and cross functional interventions will take on increased importance in accounting profession. Such professionals can significantly contribute towards organization’s strategic thinking collaborating with other parts of the organization.”
3. Potential Application of AI in accounting
- Book keeping: Book keeping is the most routine and laborious accounting activity susceptible to automation. The double entry system logic enables the coding of accounting entries. The complex accounting transactions can be expressed in accounting terms and entered into the ledgers. The accounting tasks can be automated using AI to improve accounting accuracy.
- Fraud prevention and detection: AI can be used for fraud prevention and detection as computers driven by predetermined rules cannot be enticed. The malicious activities like asset embezzlement, tax evasion, cash skimming and larceny, financial statement falsification can be traced with AI.
- Sales forecasting: The sales forecasting accuracy is crucial for preparing operations budget and AI can improve forecasting accuracy during uncertain and risky environment.
- Big data acumen: The big data provides new acumen to managers enabling to take better decisions, crafting tactical solutions, valuation of data assets and management of risk. If accountants collect and analyze structured and unstructured data accurately, it would become a great support system for decision making and redefining business strategies.
- Predictive models & Outlier detection: Accountants can use AI based big data and predictive models to improve budgeting and forecasting accuracy. They can improve internal control and risk management by using advanced outlier detection analysis and improve the efficacy facet of auditing by analysis available data sets.
- Prudence, skepticism & communication: Accountants can effectively apply their natural prudence and skepticism for improving the quality of data and data testing. Accountant with strong theoretical knowledge, practical skills and communication skills such as presentation skills, credibility, confidence, understanding people’s view point, critical thinking can acquire significant importance in the AI age.
- Leadership skills: The leadership skills like strategic thinking, coaching, mentorship, morality and cross functional interventions will take on increased importance in accounting profession. Such professionals can significantly contribute towards organization’s strategic thinking collaborating with other parts of the organization.
- Business awareness & numeracy: Equipped with data analytics skills and complemented by their inherent business awareness and strong numeracy skills, accounts will become more valuable beyond makers of historical financial statements across organisational boundaries.
- Hybrid professional role: The accountants can participate in training or testing models, auditing algorithms in various projects and integrate results into business processes, handling outliers and preparing data. In the future accounting profession will be become of hybrid nature due to the interaction of finance, analytics and AI capabilities.
- Internal audit & Expert systems in taxation: AI can be used for sorting and examining business transactions susceptible to fraud during internal audit. AI expert systems can be applied for authorizing and processing various claims, cash-flow examinations, evaluation of merger and acquisitions and investment decisions, calculating financial ratios and preparation of financial reports for filing with regulators. Various expert systems are used in the tax area such as tax treatment on stocks investments, assisting in the work of corporate tax accrual and planning process, calculating value added tax, and international tax planning and optimizing international corporation tax position.
“Various expert systems are used in the tax area such as tax treatment on stocks investments, assisting in the work of corporate tax accrual and planning process, calculating value added tax, and international tax planning and optimizing international corporation tax position.”
Way-Forward
The strong AI revolution is bringing significant changes in the economics, finance and accounting profession’s role and functions. The advent of artificial intelligence is a single most important event in the history of economics, finance and accounting and objective should be set to effectively harness the power of AI to enhance the economic conditions in which we all can flourish. In this direction following may be considered for harnessing AI in the best interest of mankind.
- An exhaustive listing of the accounting activities may be done to ascertain activities most receptive to automation.
- The accountants should be sufficiently equipped with a proper skill set to work in AI ecosystem to derive benefit from the technology advances in a volatile business environment and changing management requirements.
- The right approach to deal with the challenge of job displacement is to analyse which accounting activities can be substituted by AI, when and how. Accountants can play higher level roles in strategic areas.
- Such skills can be instill through suitable education and training. The continuous and lifelong training and learning is the key for successful adaptation to the ever changing competency requirements.
It is important to espouse the technological wherewithal and adapt to the new business and management requirements by developing competencies in AI. The future lies in AI and we must prepare for it.