What are the ethical issues that accountants should be aware of when using artificial intelligence?

THE TERM "artificial intelligence" (AI) has been in existence since 1956. The recent burst of publicity regarding AI began in 2022 with the release of Generative AI (GAI) technologies. For the accounting profession, GAI’s strength is data collection, data analytics and report writing.  

The Implications of AI 

The accounting profession and its members are struggling to grasp the implications of AI. The bedrock of AI is controversial; the collection of data is built on the abuse of copyright laws. However as outlined below there are a number of additional risks to consider. 

Unique Risks of GAI 

Since the finance department is charged with safeguarding the organization’s assets, the following questions should be asked by the accountant:

Governance. Who in the organization is responsible for AI? Do the Board of Directors and the Audit Committee understand technology risks such as GAI, cyber threats, cloud computing and blockchain? How will the organization formulate policies to address the appropriate use of GAI? 

Regulatory. How will the organization comply with the laws and regulations governing GAI? This is a particular challenge given how quickly these requirements are evolving. 

Business Case for GAI. Has the organization identified appropriate processes where GAI can add value? Has the organization defined parameters on the application of GAI? 

Selection of AI Technology. Does GAI correspond with the company’s strategic plan and information technology environment? If GAI is outsourced to another entity does this third party provide a service auditors’ report? How easily can the GAI software be modified to suit the unique requirements of the organization? 

Staff Competency. Can the organization hire competent employees who understand GAI? How are staff trained in order to maintain their understanding of GAI? 

Fraud. GAI creates additional fraud risks such as the ability to create fictitious vendors or inflate revenues. How will the internal controls adapt in order to prevent, or detect in a timely manner, the unique risk profile of GAI?  

Data Privacy. What sensitive data is GAI using, or creating, and what controls are in place to protect this data? 

Prompts. How is access to GAI restricted so that individuals cannot initiate unauthorized prompts to the software? 

Security. Does GAI increase the vulnerability to cyber attack? What additional controls are required in order to prevent data poisoning or malicious prompts? 

Model Performance.  Is the GAI technology periodically evaluated to ensure it continues to add value? How will hallucinations, that can lead to unreliable results, be identified? What independent sources have been identified in order that the results from GAI can be verified?        

How to Address These Risks 

It is this last point, model performance, where the accountant can add the most value. A unique skill that the accountant possesses is the ability to reconcile. For example, the bank reconciliation and inventory counts verify that the balances in the general ledger agrees to third party data. 

With the volume and complexity of data generated from GAI it is essential that this data can be independently verified. Two examples of the efficacy of detective controls follow. Although these examples are not AI systems these cases demonstrate the importance of confirming data to independent sources. 

The largest fraud in corporate America, the Bernie Madoff Ponzi scheme, was detected by the analyst Harry Markopolos. Mr. Markopolos, who worked for a firm that competed with Bernard Madoff Investment Securities LLC, reconciled the financial returns of Madoff securities to the stock market returns. Mr. Markopolos realized that these returns were inflated to such an extent that there was no possible explanation other than fraudulent reporting. He uncovered this fraud nine years before it was eventually acknowledged by the U.S. Securities and Exchange Commission. 

A more recent example is the sub-postmaster scandal in Britain. This scandal originated with errors in the software program used by the Post Office in the UK. The Post Office concluded, despite may indications to the contrary, that its sub-postmasters were stealing money. Almost 900 sub-postmasters between 1999 and 2015 were convicted of fraud. Many lost their wealth, several served time in jail and 15 people committed suicide.

What is particularly upsetting about these cases was the role of the judges. An important attribute of adjudicating a case is to seek independent evidence that supports or refutes the charges. The prevailing mindset of most of the judges was along the lines “if the system says that this is the number then it must be right”. Had the judges requested evidence of wrongdoing then it would have become obvious that the funds were not stolen and that the fault was with the software program. 

In regards to GAI, identify where the GAI results are being used to generate financial data. From there it is simply a matter of reconciling the data to a third party source. 

Conclusion 

The accountant has an essential role in the artificial intelligence new wave. It is up to our profession to speak up and take charge of this new technology.  

Philip Maguire, CPA, CA, is a principal in Glenidan Consultancy Ltd. His practice focuses on internal controls over financial reporting for a number of publicly listed companies on the Toronto Stock Exchange. Philip teaches a number of CPD (continuing professional development) courses in Canada, England & Wales and Ireland.

Title image: iStock photo ID:1812733266. Author photo courtesy Philip Maguire.