December 2024
Artificial intelligence or AI is the buzzword in today’s business world. The big tech companies are racing forward in providing the best of the best competitive solutions. Many are wondering what jobs AI will replace and how AI will impact the accounting industry. First, let us define what exactly AI is, how it works, and who is in the game.
According to the annual research project conducted as of 2024, Artificial Intelligence Index Report, conducted by Stanford University, AI offers “a wide range of AI capabilities, including language processing, coding, computer vision (image and video analysis), reasoning, audio processing, autonomous agents, robotics, and reinforcement learning” (p. 76). According to the study, the top indicators for capability and performance trends during 2023 included:
Outperforming human performance in certain areas, such as “visual reasoning” and grammar comprehension in the fundamental subjects of math and English.
The tech industry is leading advancement while academia comprised about 25% and academic industry partnerships contributed an additional 24% of developing machine learning models during 2023.
Cost of development is substantial in the millions reaching nearly $300 million just between OpenAI GPT at $78 million and Google Gemini Ultra at $191 million in private investment yet public funding is in the billions.
The US led the industry and outperformed China and Europe in producing over 50% of the state-of-the-art leading models during 2023.
Leaders in the industry included: “OpenAI, Anthropic, Hugging Face, and Inflection” (p. 5).
AI improved production, work performance, and communication between ability levels of workers when carefully managed.
Thus, the trend in evolution is moving to industry dominance over initial academic research due to the availability of resources. However, governments are getting involved with funding educational research through regulation due to concerns about AI being used productively and ethically managing risks that concern everyone. While AI outperformed humans in some areas, humans still outdo computers in many areas and AI is now integrated economically and functionally into most industry and government including education, science, healthcare, the economy and of course technology (Stanford University, 2024).
Human intervention continues to prevail as concerns around various risks continue to grow. For example,
With generative models producing high-quality text, images, and more,
benchmarking has slowly started shifting toward incorporating human evaluations like the Chatbot Arena Leaderboard rather than computerized rankings like ImageNet or SQuAD. Public sentiment about AI is becoming an increasingly important consideration in tracking AI progress (Stanford University, 2024).
Furthermore, researchers are concerned about many problems that need to be addressed and future problems that may arise. Some scholars are concerned about running out of data and others are concerned about the resulting potential issues with the response of computer-generated data. Some test results even indicated that AI generated data caused differing algorithm failures in poor quality data and even a loss of data. Some of these concerns raise interesting questions:
1. How can we run out of data since new data accumulates throughout time?
2. Can a computer make ethical decisions we can trust?
3. How will potential loss of data impact financial markets or accounting records?
4. Can we trust computers to dictate health or medical decisions?
Regardless, AI is evolving and here to stay and human intervention is obviously a critical component necessary in evolutionary AI. Deloitte claims AI is not superior to human knowledge and understanding needed in decision making as AI has significantly improved visual data needed to make better and faster decisions, human oversight and requires a “driver at the wheel” (Cassidy, 2024, para. 5). Forbes reported in another study machine learning performed well in auditing but lacked perfection in the areas of tax accounting and financial decision making (Marks, 2024).
Thus, human intelligence continues to dominate accounting and professional disciplines currently. AI prompts many questions about ethics and moral experience needed to sustain humanity. Future accountants should expect to have enhanced technology that produces complex financial data readily available. Accounting professionals are anticipating excellent advancements enhancing internal controls. Accounting students should be prepared to make critical analyses using advanced data analytics in real time which is why Purdue has added these types of programs in accounting since AI is here to stay!
Sources:
Cassidy, B. (2024, April 8). Empowering accounting professionals: The transformative role of Generative AI in accounting and financial reporting. The Pulse. https://rb.gy/v8oybe
Marks, G. (2024, Jan 1). The (Very) emerging role of AI in the accounting industry. Forbes. https://rb.gy/05o76c
Stanford University. (2024). Artificial Intelligence Index Report 2024. https://rb.gy/x2m7ay
Sylvia DeAngelo, Ph.D.
Accounting Department
https://aiindex.stanford.edu/wp-content/uploads/2024/05/HAI_AI-Index-Report-2024.pdf
Investigative Accounting Series
by Cynthia Waddell, PhD, CPA, CFE
Asset Misappropriation – Payroll Fraud
In previous months, we looked at the fraud theories that forensic accountants use to analyze fraud. In this article, we will examine how forensic accountants can assist in asset misappropriation fraud schemes (taking an employer’s assets without authorization), specifically, a payroll fraud case. Payroll fraud falls under the category of cash disbursement fraud schemes.
A valuable tool that forensic accountants employ is data analytics. Some basic data analytics techniques that forensic accountants use to investigate asset misappropriation fraud schemes are duplicate testing, even dollar testing, gap tests, join/relate, ratio analysis, threshold testing, and trend analysis.
Ghost employee payroll fraud schemes are often not complex and can be perpetrated by just one person. The fraud that occurred at a Wisconsin health care facility is one such example. From 2013 through 2017, a payroll executive embezzled over $476,000 from her employer using a ghost employee scheme (Vielmetti, 2020). The executive created a fictitious male employee and deposited payroll checks for this ghost employee in her own bank account.
Although there was an obvious lack of segregation of duties at this company, simple data analytics techniques could have been used to search for duplicate addresses, phone numbers, Social Security numbers, and bank account numbers. The forensic accountant could also perform gap tests – searches for details of personal information that is missing, such as emergency contact, telephone number, or bank account, that a legitimate employee would typically provide to the employer. Another test would be to match employees in the payroll register with employees in the human resource database, as well as verifying the employee has a performance evaluation record. If the company has a security database, is the employee included in the database? The auditor could also search for changes to the payroll master file, for example, to identify terminated employees who appear as active employees.
The forensic accountant could also conduct additional analyses of the payroll records. The forensic accountant could search for employees who have no voluntary deductions from their paycheck or have a high ratio of net to gross pay. A few basic data analytics tests could have raised suspicions about this employee and greatly shortened the length of the fraud!
Case Reference
Vielmetti, B. (2020) Payroll executive charged with embezzling more than $476,000 from Glendale firm. Milwaukee Journal Sentinel. Retrieved April 19, 2021, from https://www.jsonline.com/story/news/crime/2020/01/13/payroll-manager-used-fake-employee-steal-her-boss-charges-say/4434988002/
Cynthia Waddell, PhD, CPA, CFE