A recent tweet (or whatever they’re called now) seemed to sweep our whole profession into the modern technology-facilitated trash. It said: “What do the IFRS people know about running a bank? You don’t need such people to make financial statements – Chat GPT is there for such things…”

Well, as I wrote here, ChatGPT itself isn’t quite there yet. But there’s certainly a lot of talk about how artificial intelligence and enhanced automation might affect financial reporting more broadly. A recent LinkedIn post by Shawna Johnson covers some of the more commonly-cited areas:

• Traditionally, financial professionals would spend a significant amount of time gathering and organizing data from multiple sources. With automation, this process has become more streamlined, with tools that can automatically retrieve and consolidate data from various systems, such as accounting software, bank statements, and market databases.
• Automated financial statement generation is another area where automation shines. By leveraging pre-defined templates and algorithms, organizations can easily generate financial reports. This eliminates the need for manual data entry and reduces the likelihood of errors. Furthermore, automation allows for real-time updates, ensuring that financial reports are always up to date.
• The use of automation in financial reporting also improves accuracy and reduces errors. The chances of human errors, such as typos or miscalculations, are minimized by removing manual intervention. Additionally, automation can perform complex calculations and cross-verification processes, ensuring accuracy and consistency in financial reports.
• Apart from efficiency and accuracy gains, automation in financial reporting brings significant time and cost savings. Tasks that previously required hours or even days can now be completed in minutes. This enables financial professionals to allocate their time and resources to more value-added activities, such as data analysis and strategic decision-making.

It’s perhaps even easier to see how AI might reduce some of the standard audit-related drudgery. A recent KPMG article covered some of this:

• We expect AI will continue to help KPMG firms to better identify high risk transactions, allowing us to sustain our focus on risk assessment and obtain audit evidence over much larger, more complex sets of data. Also, by removing some of the more time-consuming tasks for auditors, AI will free them up to apply valuable skills in other areas, again enhancing the audit for everyone.

• … KPMG has always believed that technology, including AI, has the potential to empower and enhance auditors in their day-to-day work and provide real value to an audit. But, we must never simply rely on AI as a black box… That’s why we believe AI will never replace people and KPMG will always have human knowledge in the audit loop.

KPMG note that “A key risk around AI, which is very relevant to audit, is ‘explainability’. By this we mean we still need to understand and explain why the technology may be highlighting certain items and trends. That’s why we believe it’s important to have people working alongside AI to prompt deeper thinking and challenge where necessary, rather than removing people from the loop entirely. At the same time, this will help training the AI continuously through human input.”

(The Ontario Securities Commission also recently issued a report on AI in Ontario’s capital markets, but the focus there wasn’t on financial reporting: areas addressed include asset allocation, price and liquidity forecasting, hedging, trade order execution and surveillance, high-frequency trading, futures market analysis, and sales and marketing).

It’s amusing how both the above extracts trot out the time-honored notion of freeing people from mundanity, allowing them to spend their time in more “value-added” areas. The same was said about computerization, the Internet, Blockchain…for all I know they were saying it in the distant past about the wonders of ballpoint versus fountain pens. How can it be, you might reasonably ask, that we never run out of boringly time-consuming tasks from which humans need to freed? But of course, like all those previous advances, AI is just as likely to create previously unimagined kinds of non-value-added activities (no doubt, for instance, the “explainability” challenge summarized by KPMG will eat up more human time than currently envisaged), or to trigger human resource cuts that leave the remaining staff members no better off. All of which may only add to the declining interest in joining the accounting profession…

I don’t tend to subscribe to dystopian technological fantasy; like KPMG I don’t envisage people being entirely supplanted by the machines they created. And yet, financial reporting might be more susceptible than many areas and functions to a version of that: a swarm of data prepared by AI, with audit-quality verification built into the process, and delivered directly to AI users who adjust their investment and other models accordingly, all of this existing in perpetual equilibrium with no human input required. Just think of all those fortunate displaced would-have-been-accountants, set free to pursue lives of undiluted value-added activity!

The opinions expressed are solely those of the author.