This article by D.J. Marini—Founding Partner of Burkland’s Accounting Practice—was originally published on Medium.
Artificial Intelligence (AI) is transforming multiple industries, revolutionizing everything from human resources to legal services to accounting. Organizations already are using AI to save time and scale efforts in all aspects of business, such as creating marketing content, crafting computer code, helping with hiring, and/or chatbots for customer service. A 2023 survey from research firm McKinsey indicated one-third of respondents were using AI already in at least one business function. The report also said 60 percent of companies with existing AI capabilities are employing AI tools, but only 21 percent of organizations say they have policies in place to manage AI’s usage.
The accounting industry is particularly primed to reap the benefits of AI, as these tools can help professionals make their workflow more efficient and work product more effective. Mordor Intelligence revealed in a recent report that the market size for AI in accounting is currently at $4.73 billion, and it’s expected to hit $26.66 billion by 2029. There is tremendous potential to analyze vast amounts of data and produce key insights quickly. However, as with any technology, there are risks as well. While embracing the power of AI, there are potential pitfalls to avoid, and human oversight will always be a must.
For any organizations considering applying AI to the accounting practice, the following aspects should be kept in mind.
AI can automate mundane accounting tasks
Accounting is compliance-focused, so AI can take the repetitive tasks accountants must perform, including data entry, reconciliation, and invoicing. Automation of these potentially helps reduce human error, while allowing professionals to focus on more valuable and nuanced work such as strategic financial planning. And predictive analytics allow AI to identify risks, recommend remediation steps, and even foresee potential violations. AI can see the entire story that the data is displaying, making it even better able than humans to see patterns and identify discrepancies. Financial reporting accuracy is improved, and real time insights can be garnered.
However, there is also a risk regarding transparency. AI tools might leverage opaque algorithms to accomplish data analysis that make it challenging for humans to understand. And not being able to see how AI arrived at a given decision makes it difficult to judge it.
AI can affect privacy and security matters
It’s critical to be careful in sharing sensitive information, financial or otherwise, with AI tools—particularly if entering information into an external platform. Nonproprietary AI tools may retain and reuse inputs, where confidential data ends up being inadvertently shared with third parties. This situation happened to Samsung when the company discovered staff had “leaked internal source code by uploading it to ChatGPT.” And accounting professionals are dealing directly with information that must be kept safe and secure.
If AI has access to sensitive financial data like social security numbers, bank accounts, and tax info, there’s risk of identity theft and more. It may be stating the obvious, but accountants should only use a trusted tool that offers clear data protection capabilities like advanced encryption and security protocols.
AI can help with research, planning, and training
Research can take time, and AI can help minimize the amount needed. Also, laws and regulations are always evolving, and AI can assist with laying out multiple, possible scenarios—exploring the industry, trends, and overall landscape. Companies can gain an advantage through informed, real time, industry-based decision making that helps develop sound strategies and better allocate resources. AI can also assist in understanding customers. By reviewing transactional data and combining this with current and predicted market conditions, future outcomes can be intelligently suggested.
However, a company still must have human review, and training to support that review is important.
AI can aid but not replace humans
In complex situations or building client relationships, AI cannot act alone. While these tools can help detect fraud or reveal patterns, they’re not as useful in understanding a complex or unusual situation such as a merger. There’s an incredible amount of nuance in a situation like an acquisition, where a human was involved and needs to be in subsequent analysis.
With clients, where it may be possible to draft a response based on prompts, saving some time, people ultimately want to deal with people. If a client knows or even suspects they aren’t interacting with a human in key communications, trust is potentially lost. And with chatbots, where there is benefit in delivering customer service and support anytime anywhere, there is also the risk of frustration if the AI system is not sufficiently advanced to effectively handle queries.
AI can save time, but it isn’t an unimpeachable shortcut
It’s important to always check any and all AI-generated information—people still play a pivotal role in confirming, cross-referencing, and verifying results. Many AI tools were “trained” on protected data sources, meaning there’s some legal uncertainty whether AI tools encroach on intellectual property law. AI may accidentally replicate data from these external sources, so plagiarism detection might be necessary as well to adhere to ethical standards. A financial report, for example, could easily and inadvertently leverage another company’s report.
Accounting professionals and organizations must ensure it’s clear how information gleaned from AI was arrived at and that it’s accurate. Also, AI can produce biased content—biases may be present in original datasets used for training, or AI may introduce bias according to its own data interpretations. AI tools can process information incredibly fast, but often without nuance, understanding, experience, and ethical judgment.
AI’s promise has been recognized, and again, it’s already being incorporated into a number of job functions. But the ongoing need for human oversight cannot be sufficiently stressed. AI can assist with certain accounting needs such as bookkeeping, tax preparation, and insights, but it can’t replace an experienced accountant who has the expertise to review the data and recommend decisions.