Business Studies
Topic: State of Knowledge of Artifical Intelligence
Current State of Knowledge of Artificial Intelligence
Introduction
Artificial intelligence (AI) is the representation of human intelligence in machines that are programmed to think like and copy the actions of humans. They are machine portraying characteristics associated with the mind of humans, such as problem-solving and learning. AI is a technological advancement that is changing the model of operations in organizations. It already has considerable effects on network marketing services and processes through Assessment of the behavior of users in networks and development of user-profiles in which service and product offering are focused. AI impacts production departments by automating quality control, detecting abnormalities in lines of production before the occurrence of a problem, and managing maintenance in a predictive manner. Similarly, AI impacts logistics processes by maintaining contacts with the providers of logistics service and clients in an automatic manner. It also affects after-sale services by evaluating the opinions of clients concerning services or products to measure their satisfaction level and possible improvement or failure that may apply to them. Therefore, the knowledge of artificial intelligence has increased in the current years, which has led to its incorporation into companies.
For more than two decades, technological innovations have been the essential drivers of economic growth. The most vital technological innovations are what is referred to by economists as multi-purpose technologies (a group comprising of the internal combustion engine, electricity, and steam engine). Each of these technologies increased waves of complementary opportunities and innovation. For instance, the internal combustion engine has led to the production of airplanes, cars, trucks, and other vehicles. Diverse companies such as UPS, Uber, and Walmart have devised new ways to influence technology to produce new business models that are highly profitable. They use AI technologies such as chatbots and google maps to communicate with their customers and find directions in different parts of the country.
AI is the essential modern-day multi-purpose technology. It comprises machine learning (ML). That is the ability of machines to keep enhancing their performance with limited human interventions. Humans have little understanding of how the machine successfully achieves all the tasks it is given (Brynjolfsson & Mcafee, 2017, 3). Machine learning has become widely available and far much effective within the past few years. Systems and machines that learn how to carry out a particular task have been developed. Before the arrival of ML, humans were not able to articulate their understanding to automate several tasks. People in the current world make use of ML technologies in their daily operations.
Artificial intelligence’s original goal was to change non-analytical knowledge of humans into computational information from the symbolic process of computation, connectionist networks, or a combination of the two. AI is therefore considered as science with double analytic lines, theoretical and correlation (Brynjolfsson & Mcafee, 2017, 6). The correlation describes the successive and simultaneous of different inferential procedures to which symbolic scales are assigned. The theoretical analysis seeks explanation based on a number of guidelines in a higher knowledge plane that contributes to predicting and describing future and past events.
Nowadays, artificial intelligence has been redirected towards solution construction to challenges with huge data volumes, which change after some time, presenting contradictions and inaccuracies. The basic change since the development of AI in the ’60, the ’70s, and 80s, and its rigorous, intensive use in the current century is the move from the symbolic, connectionist approach (Ruiz-Real, Uribe-Toril, Torres, & De Pablo, 2021, 100). Therefore, the skilled systems based on guidelines and rules were the traditional AI’s spearhead for problem resolution in areas like geology, medical fields, recovery of data by use of natural languages, operation planning for robots, and chemical analysis.
Importance of Artificial Intelligence in Business
Nowadays, companies incorporate AI technologies to increase efficiency, enhance customer service, reduce operational costs, and grow revenue. To attain higher returns, companies should consider putting the full modern technologies such as natural language processing, machine learning, and many others into their products and services. Some of the major benefits of AI in business include saving money and time. Machines are more efficient than humans when it comes to working (Lin, 2020, 1). AI machines can operate the whole day and night without getting tired or bored. Machines will neither ask for breaks nor time to sleep. Thus, they are reliable anytime an individual wants to operate them. AI machines are programmed to notify the operator about any production fault or any significant event reducing losses that may occur. Similarly, AI machines can examine a large amount of information for a short time, reducing time wastage. They often take a shorter time in decision-making so long as they have significant information.
Besides, AI enhance customer experience. Artificial intelligence machines allow firms to offer customer service at any time of the day or night. They have been used mostly in the hospitality industry to improve customer service (Ruiz-Real et al., 2021, 112). The companies such as Euro Disney operate with one single website app to offer customer service. The client will log into the app to report any complaint then the app operator directs the message to relevant personnel to handle it. Many customers can interact simultaneously, and their questions responded to through the website app. Besides, AI has enabled organizations to easily communicate telephone calls, emails, and even online chats. Therefore, it helps enhance customer service.
Furthermore, artificial intelligence machines minimize errors in the company. Even though they are not free of error, they are more accurate than humans. Their accuracy frequently 99 percent and above. For many companies, cash flow forecasting has been the process prone to error and most time-consuming (Lin, 2020, 2). AI helps in enhancing accuracy in cash flow forecasting with less manual interference. Businesses are therefore given higher chances of attaining success through artificial intelligence.
Moreover, artificial intelligence has several benefits in a medical setup. Smart gadgets are essential in the medical environment for patient monitoring. Employing AI to improve the ability of identifying deterioration or to detect development of complication in a patient can significantly enhance outcomes and minimize costs that may arise due to penalties from hospital-acquired conditions. AI has enabled medical practitioners to integrate information across the healthcare system. It generates alerts that signal an ICU physician to intervene early (Brynjolfsson & Mcafee, 2017, 8). Therefore, AI minimizes preventable risks and deaths in the hospital. Adding intelligent algorithms into hospital devices can minimize reasoning stress from doctors while making sure that patients get the required care on time.
Another importance of AI is that it promotes easy monitoring of an individual’s health through personal devices or wearables. In the modern world, almost every individual access to gadgets with sensors that gather essential information about their health. Some of the monitoring devices are connected to smartphones that give the owner alerts of abnormality in their health data. Some are wearables that can measure the pulse rate breathing rates. Gathering and analyzing information through wearables and personal devices when supplemented with data provided through home monitoring devices and other apps is important since it ensures a doctor’s early intervention when an illness occurs. Through AI devices, many lives have been saved in the past few years.
Limitations of Artificial Intelligence
Despite bearing multiple advantages to both the business and individuals, AI has some limitations. The first disadvantage is high installation and maintenance costs. AI machines are very expensive to purchase; the company uses a large amount of money to buy a single machine. The creation of AI technologies is costly due to their complexity, so they are expensive. Furthermore, it requires highly skilled personnel to install. AI machines software and hardware need to be updated with time to achieve the latest requirement and ensure smooth operation. Machines require regular maintenance and repair, which need a lot of capital. The company will have to employ skilled personnel to monitor and maintain machines regularly. Similarly, the existing employees will require training on how to operate and handle AI machines. The pieces of training are costly, which raises the general operating cost of AI machines.
Besides, the incorporation of AI machines into a business increases unemployment rate. A single AI machine can perform multiple tasks which can be carried out by several human employees. The incorporation of machines replaces human labor. Organizations nowadays are striving to purchase robots and machines to carry out tasks with more efficiency than minimum qualified individuals (Prevedello, Halabi, Shih, Wu, Kohli, Chokshi, Erickson, Kalpathy-Cramer, Andriole, & Flanders, 2019, 11). A single individual only operates machines; therefore, the number of employees in an organization with AI machines will be minimal. When the employment rate is low in a particular country, the living standards also reduce. Poverty rates also increase with an increase in the rate of unemployment.
Moreover, AI machines cannot multi-task. AI machines can only carry out tasks which they are programmed and designed to perform. They will crush or give irrelevant output if they are given tasks beyond their description. The machine cannot think outside the box. Therefore, giving it a task that is not described for it may lead to its failure. The company thus will have to incur the high cost of repair.
AI promotes laziness. The automation of most of the tasks in the machine, it is making humans lazy. Humans will entirely depend on a machine to perform their tasks, for instance, with the introduction of automatic vehicles, which do almost everything for the driver. The only major task the driver has is to control the steering will and to step on the gas (Prevedello et al., 2019, 10). Gear changing has been automated in these vehicles. Too much comfort and less task in these vehicles may induce laziness while driving, and the driver may be tempted to sleep. Humans are always addicted to the automation of tasks which may create challenges for the future generation.
Finally, AI has led to a reduction in team building in companies. Lack of team building is mainly because machines have no emotions. It is clear they work much better and efficiently, but they cannot replace the creation of a team by humans due to their connection. Team building is all about bond development among humans. Machines lack that vital feature when it concerns team building.
Challenges of Artificial Intelligence Machines
With several organizations expecting that AI use can enhance business productivity up to 40%, the increase in the AI number has significantly increased since 2000. AI applications can range from pursuing asteroids and other cosmic bodies in the universe to disease prediction on earth, explore innovative and new ways of curbing terrorism, to develop industrial designs. It has so many challenges despite its growth and achievements. Its main challenges include a trust deficit. One of the most aspects that causes worry for artificial intelligence is the unknown manner in which the learning models forecast output (Prevedello et al, 2019, 15). How a particular set of inputs may produce results and solutions for various problems is not easy to understand for humans with less expertise. Lack of knowledge and understanding has led to a reduction in trust for AI machines. Several individuals around the globe do not even know the existence or use of AI and how it is connected with daily devices they interact with like smart TVs, automatic cars, banking, and smartphones.
Another challenge is the little knowledge about AI. Even though there are several places people can learn the existence and use of AI, there is still limited knowledge about it. Apart from researchers, technology enthusiasts, and college students, few people understand AI’s full potential. For instance, several medium and small enterprises can have their work learn or schedule innovative means of enhancing their production, managing and selling products online, managing resources, studying and understanding consumers’ behavior, and reacting to the market efficiently and effectively (Prashanthi, Deva, Vadapalli, & Das, 2020, 4). They also do not understand the use and existence of service providers like Amazon web services, Google Cloud, and other industries for technology.
Besides, human level is another important challenge facing artificial intelligence. It has kept the researchers on their toes to investigate AI services in startups and companies. AI machines can provide accuracy of 90 percent and above, but humans perform better in all these setups. For instance, if a machine is given a chance to predict whether an image is a cat or a dog. Nearly every time, humans can predict a correct output, attaining an accuracy of over 99 percent. For an ML model to achieve the same performance, it would need hyperparameter optimization, extraordinary finetuning, an accurate and well-defined algorithm, and a large dataset, supplemented with vigorous computing power. That is a lot of work, and it is more difficult.
Moreover, the lack of security and privacy of data is a significant challenge of AI. The primary factor on which machine learning and deep models are grounded in the resource and data availability for their training. Due to increase in cyber-attack, the data can be used for the wrong purpose since they are generated for several people in the world (Prashanthi, Deva, Vadapalli, & Das, 2020, 2). For instance, if a medical practitioner has provided services to more than one million individuals in a town, but due to cyber-attack, there may be some cases of data leakage. The patients’ data comprise information about medical history, health problems, disease, and much more. Leakage of the patients’ data reduces their privacy.
Artificial Intelligence in Finance
Machine learning and artificial intelligence in finance incorporate everything from fraud detection, automation of tasks, and chatbot Helpance. Currently, a considerable number of banks understand the possible benefits of AI in finance. The financial institutions’ decision to embrace artificial intelligence is enhanced by advancement in technology, shifting frameworks for regulatory, and increased user acceptance. Financial institutions that are already using artificial intelligence are capable of streamlining tiresome processes and greatly enhance customer experience by providing whole day and night financial advice services and access to their accounts. Besides, AI in finance can drive efficiencies in areas ranging from trading and managing risks to claims and underwriting. Certain applications are relevant to a particular financial service division, while others can be forced through the board. Artificial intelligence in finance is useful in the following sectors:
Artificial Intelligence in Financial Risk Management
When it comes to the detection of fraud and security, AI has been proven to be highly valuable. Conventional ways of detecting fraud comprised of examination of information and data by computer against certain rules and guidelines. For instance, a particular financial institution set a maximum day transaction to be 20000 dollars in that any transaction beyond the set threshold would be highlighted for further examination (Lin, 2020, 1). Nevertheless, the traditional analysis required additional effort and produced several false positives. Sometimes, cybercriminals would often change their strategies. Thus, the most efficient and effective risk management system must continue to evolve and become smarter. With enhanced learning algorithms like deep and machine learning, new features can be incorporated into the system for vigorous adjustments. With cognitive analytics, the models for detecting fraud can become more accurate and robust. If an AI machine rejects what it verifies as possible fraud and humans considered it a non-fraud element due to certain reasons, the AI technology will learn from humans’ insights, and next time it determines that similar element, it will not reject it. Artificial intelligence machines are getting smarter.
Besides, AI technologies have enabled financial institutions to process vast data without any loss due to fraud. For instance, the biggest payment institution in the entire world, PayPal, and its enhanced guidelines. The company has a huge target on its back due to its visibility and scale. In 2015 alone, it handled over 200 billion dollars from 4 million transactions by its numerous customers (Lin, 2020, 2). Nonetheless, PayPal company has managed to improve its security by investing in deep and machine learning technology. Currently, the fraud rate in PayPal revenues has dropped to 0.32 percent (Lin, 2020, 2). PayPal was using a simple linear style in the past. Currently, it uses algorithm data mine from the purchase history of a customer and examines the possibility of fraud patterns stored in its developing database. Deep learning can control up to a hundred thousand data points, whereas the linear model only consumes 30-20 variables. The incorporation of AI technologies and its advanced capabilities enables PayPal to differentiate frauds from innocent transactions. AI, therefore, has greatly enabled PayPal to minimize and manage financial risks.
Artificial Management in Financial Trading
Investment management firms for many years have largely depended on computers to trade. About 1400 hedge funds, comprising 9 percent of total finances, depend on large statistical styles developed by quants. Nonetheless, the statistical models only use historical data, require humans intervention, are always static, and the performance reduces with a change in the market. Subsequently, finances are gradually moving in the direction of true AI models that keep improving themselves and analyze huge data volumes. The AI technologies use complex techniques comprising evolutionary computation, deep learning, and Bayesian networks stimulated by genetics (Brynjolfsson & Mcafee, 2017, 11). Artificial intelligence trading software can consume huge data volumes to study the world and provide financial market predictions. The software absorbs information from news reports, tweets, books, earning numbers, and financial data to understand international market trends. Artificial intelligence trading platforms are far much different from high-frequency trading. HFT enables traders to implement over one million orders and scan several markets at the same time, therefore, reacting to opportunities in a unique way. AI-controlled trading platforms are pursuing the best trade, and strategy is dictated by machines, not humans.
Artificial Intelligence in Finance Robo Advisory
Robo advisor is a digital platform offering algorithm-driven and automated financial planning services with little human supervision. Since early 2000, investors were employing advisors to obtain returns from technologies, whereas human, financial managers were using automated portfolios. In the modern world, robo-advisor enables clients to get direct access to the service. Robo-advisors are reliable since they monitor the market throughout and are always available, unlike human advisors who need time to rest and sleep. Besides, research has shown that robo- advisors can provide a cost-saving of up to 70 percent and always need minimum human participation (Lin, 2020, 4). Currently, robo-advisors can Help in performing multiple tasks such as asset transfer and account opening. The process only requires the customer to respond to a simple questionnaire about liquidity factors and risk appetite, translating to investment judgement by robo-advisors. Many modern robo-advisors aim to assign to the managed portfolio based on their choices. In the future, it is believed that the capabilities of robo-advisors will advance into more developed contributions like automatic asset shifts and enlarged coverage all through asset classes such as real estate.
Besides, Robo-advisors have an impact on personal wealth and financial management division. Currently, total robo-advisory assets under management only comprise ten billion dollars of the wealth management industry which is four trillion dollars. The robo-advisory assets represent less than one percent of all account assets managed. Research has pointed out that by the end of 2021, the total assets under management will rise to ten percent (Lin, 2020, 4). Therefore, personal wealth and financial management will be greatly enhanced.
Moreover, the players in the financial industry have embraced different robo-advisory approaches. Smaller management companies incorporate components of algorithms to automate the management of investments, compete with robo-advisory, and reduce costs. Well-established financial institutions are purchasing already existing robo advisors like the acquisition of Jemstem by Invesco or develop their own solutions of robo-advisors like Intelligent Advisory of Schwab and FidelityGo.
Insurance Underwriting and Claim
Insurance mainly relies on the risk balance among different people; insurance companies group together people with financial similarities in which some will need payouts while others will not. The insurance industry is based on risk Assessment and assessment; they are not strangers to data analysis (Corea, 2019, 6). Nonetheless, artificial intelligence can enlarge the volumes of data analyzed and the ways they can be used, leading to more operational efficiencies and accurate pricing. Startups are at the front line in propelling the insurance industry. They show what is possible and what can be done.
Artificial intelligence can enhance the automation of a significant underwriting amount, particularly in well-established markets with data availability. Today, the underwriter of insurance, with Helpance from actuarial models and computer software, evaluates the exposure of potential clients and the risks associated, how much they should be charged for coverage, and how much they should receive. Artificial intelligence can help automate large amounts of underwriting into home commercial, group, home, and life insurance (Corea, 2020, 9). AI will improve modeling in the future, emphasizing major considerations for decision-makers that may have gone without being noticed by an insurance company. Besides, it is also expected that improved artificial intelligence will permit personalized underwriting by an individual or a company, considering circumstances and behaviors.
Besides, improved underwriting will influence machine learning for mining data, deep learning facial analyzers, and wearable technology. For instance, a startup insurance company, Laptus, wants to use selfies to foresee life expectancy correctly. In the Laptus model, clients will email their self-portraits, which will be analyzed and scanned by computers (Lin, 2020, 3). Computers will analyze several regions of the face. The analysis will put into consideration every aspect starting from simple demographics to how fast a person will age, the index of their body mass, and whether they are into drugs. Furthermore, the process of underwriting could be made more collaborative with wearable technology. Rather than depending on complicated processes of contracts and lengthy medical checks, wearables can offer real-time visions into policyholder behavior and health. All the AI technologies such as wearables indicate that machine learning in finance is advancing.
Artificial Intelligence and Insurance Claims
Insurance claims can be referred to as formal requests for payments delivered to insurance companies. The received request is then analyzed by the company for legitimacy and payout to the insured when approved. Artificial intelligence can improve the insurance claim processes by enhancing the data accuracy of the customer. The process of insurance claim process is relatively manual (Lin, 2020, 3). That is, the incident details and information of clients are logged in manually by human agents. Manual processes can result in losses due to data errors and typo errors. Artificial intelligence technologies can enhance accuracy by minimizing manual input. Moreover, processes of claims always need human agents to match the information of data with several databases. Artificial intelligence can be used efficiently to match customer information with databases.
Besides, artificial intelligence encourages faster payout recommendations. The primary cause of consumer dissatisfaction in insurance companies is the slow cycle time. AI can minimize turnaround time by verifying the policy, and then it makes the determination on the claims and decide whether to automate payments or not (Lin, 2020, 4). Artificial intelligence technologies have the ability to analyze both unstructured data (certificates and handwritten forms) and structured data, therefore, making claim processes easier.
Conversational Banking and Customer Service
Financial institutions are making bets with their customers’ computer Helpance, referred to as chatbots. Initial chatbots versions were only able to respond to basic questions about recent transactions and spending limits. In contrast, future versions are planned to be full-service virtual Helpance that can track budgets for consumers and make payments (Lin, 2020, 3). While engaging with customers can result in considerable cost savings, straightforward number crunching is easier than human interactions. However, artificial intelligence technologies such as chatbots lack the understanding and empathy needed by a human when handling difficult financial situations and decisions. Natural language AI technology is vital for handling the wishes and concerns of customers.
Artificial Intelligence in Credit Risk Management
When combined with machine learning and artificial intelligence technologies, a comprehensive credit management system can reduce financial risks and level up decision-making processes efficiency, therefore, increasing the company’s profit (Dolgorukov, 2020, 1). Credit risks are all potential risks a bank is willing to take while lending out money. Examining the ecosystem for few months may help the credit risk manager predict shifts in economic leading to unpaid credit. Furthermore, review of client’s payment history of previous to help reduce non-performing credits.
Using traditional software, creditors risk failing in their business. Some of the disadvantages of conventional credit risk management approaches include the long waiting time between application ad issuance of a loan. Traditionally, the process may take more than 3 weeks (Dolgorukov, 2020, 1). Traditional risk managers may incur a high risk of false forecasts because of the application of credit models, which are unscientific. Besides, the performance-price ratio of the major credit scoring is always questionable.
Artificial Intelligence and Machine Learning in Credit Risk Management Tools
Machine learning and artificial intelligence have allowed the possibility of automated systems, self-driving cars, and chatbots. While using AI and ML tools in credit scoring, credit landers can examine different information on the borrower, such as economic behavior and payment history. ML and AI software enhances the banking operation accuracy and reduces the process of decision-making.
Importance of AI in Credit Risk Management
AI tools have enabled the automation of lending processes. AI can minimize staff expenses through automation of processes of credit risk management. Process automation enables effective money distribution allowing workers to focus on other tasks in the company (Dolgorukov, 2020, 14). Automation of processes also reduces employment costs as it reduces the number of employees employed to handle credit risk management.
Furthermore, artificial intelligence has led to a reduction in time for credit management. Banks and other financial institutions may take almost one month to review and validate loan applications. When using scoring software, the validation time may be reduced to few minutes. For instance, modern technologies such as GiniMachine may validate over 1000 applications in less than ten seconds (Dolgorukov, 2020, 2). Besides, artificial intelligence technologies are free of error. Conventional lending procedures do not assure low-scoring errors. Whereas AI machines can handle large volumes of data and predict patterns with minimal error.
Similarly, AI machines are incorporated with technologies that strongly detect fraud. They are equipped with tech-savvy solutions that enhance mechanisms for fraud detection. The capability of AI machines to detect fraud protects banking activities and ensures a more dependable picture for creditors (Dolgorukov, 2020, 2). Besides, artificial intelligence technologies are linked with high accuracies in predictions. Conventional risk management technologies were functioning under defined rules and regulations. Making use of ML and AI allows users to get an insightful solution that can examine several databases during the lending process.
Conclusion
In a nutshell, artificial intelligence has been widely employed in different sectors in the world. It has been employed in health, business, and even in law. It has several advantages in business, such as enhancing customer service, fraud detection, and minimizing errors in a business. Despite bearing several advantages, AI also has some challenges and limitations. Its challenges include a trust deficit, lack of knowledge about technology, and privacy and security issues. Artificial intelligence has also been employed in finance and credit risk management.
References
Brynjolfsson, E. and Mcafee, A., 2017. The business of artificial intelligence. Harvard Business Review, 7, pp.3-11.
Corea, F., 2019. How AI is changing the insurance landscape. In Applied Artificial Intelligence: Where AI Can Be Used in Business (pp. 5-10). Springer, Cham.
Dolgorukov, D., 2020. AI for Credit Risk Management: Banking and Finance. Finextra Article
Lin, M., 2020. How Artificial Intelligence Is Disrupting Finance. Review of Financial Processes
Prashanthi, G.S., Deva, A., Vadapalli, R. and Das, A.V., 2020. Automated Categorization of Systemic Disease and Duration From Electronic Medical Record System Data Using Finite-State Machine Modeling: Prospective Validation Study. JMIR Formative Research, 4(12), p.e24490.
Prevedello, L.M., Halabi, S.S., Shih, G., Wu, C.C., Kohli, M.D., Chokshi, F.H., Erickson, B.J., Kalpathy-Cramer, J., Andriole, K.P. and Flanders, A.E., 2019. Challenges related to artificial intelligence research in medical imaging and the importance of image analysis competitions. Radiology: Artificial Intelligence, 1(1), p.e180031.
Ruiz-Real, J. L., Uribe-Toril, J., Torres, J. A., & De Pablo, J. 2021. Artificial intelligence in business and economics research: trends and future. Journal of Business Economics and Management, 22(1), 98-117.
