be serious about ai, not amused

seven essential things to know.

by hitendra patil
client accounting services: the definitive success guide

unless you live under a stone, you would have heard about all the craze chatgpt has created recently. and most likely, you, too, are amused by it.

more: in “aicccountants,” hitendra patil demystifies artificial intelligence for accountants | how to offer proof to cas prospects | you don’t need a cas sales pitch | get clients to toot your cas horn | how to market cas to existing clients | five methods to show your firm’s marketable value | seven ways to market your cas offering
exclusively for pro members. log in here or 2022世界杯足球排名 today.

an artificial intelligence-powered tool, it is one of the first that allowed people to directly interact with ai in a familiar way how we use chat functions on mobile phones and websites. this ease of access and usage gave it a huge fan following. moreover, it triggered a tsunami of new ai tools emerging all over the internet.

i have been keenly researching and studying ai, from an accountant’s point of view, for the last few years. i came across one of the famous quotes about ai, which has resonated with me for some time now:

“by far, the greatest danger of artificial intelligence is that people conclude too early that they understand it.” – eliezer yudkowsky

i see this quote living in reality – for example, many people, including accounting firms, started using chatgpt to generate marketing content. but then, a startup cfo used chatgpt to build a financial planning and analysis tool.

we don’t know what we don’t know.

the lowest stage of knowledge is when we don’t know what we don’t know. when it comes to ai, we may know it exists and is fast progressing. but we don’t know how it is created, learns and works. therefore, we feel it is too technical, and one must be a programmer to understand it.

not really.

as an accountant, don’t you almost always correctly know what your accounting, tax or payroll software will do? given your input, you can fathom how the software code will work. you do not know the program code lines, but you (kind of) know what it will do. why? because it is a finite code. it does not change until a new version is rolled out. therefore, you can accurately predict how it will work. so, you are not a programmer, but you know how the software “works.”

but that is not the case with ai because it can learn on its own and hence, can keep changing itself. and when that happens, constantly, you can only chase the mirage. so it is hard to keep up with ai.

not really.

we know what we don’t know.

one can argue that all acquisition of new knowledge is at the stage when we come to know what we do not know.

1. now we know that ai learns on its own, but we do not know how.
2. an implied knowledge is that ai tools get created to learn on their own, but we do not know how they are created.
3. further implied knowledge is that we experience how ai works but do not know why.

you do not need to be a technologist or a programmer to understand the three stages above.

knowing the fundamental debits and credits and the “t” accounts is your foundation of accounting expertise; learning the basic building blocks of ai can empower you to make ai a (very) powerful tool in your hands. yes, a tool. it is not a replacement for your work, although it can take away much repetitive, defined work. yet, it can be a tool in your hands, perhaps the most powerful you have ever used. but …

only if you learn how ai is created, learns and works.

it is not difficult, although it can take significant time (even months and years). it is like starting your car in the morning and looking to travel a thousand miles without knowing the destination and without your gps.

that is why i wrote my new essential guidebook on “ai exclusively for accountants.” i coined the term “aiccountants™” in 2018. i was exploring “behind-the-scenes” ai integration into accounting software. i was learning how “robots” (software code) learn from what the user enters into the accounting software. yes, your use of software can train the software! the more i explored, the more i learned, and the more i experienced that there is a distinct path to learning about ai. i believe accountants leapfrog their learning about ai without spending days, weeks and months.

amusement, confusion and even fear about ai emerge from the lack of (adequate) understanding. it is, therefore, essential to know that:

• ai has limitations (that can produce inaccurate/dated outcomes).
• ai can have biases (that can produce results that can linger on the border of ethical and unethical work).
• ai may produce outcomes not relevant to your clients (as data used to train ai may be irrelevant to your clients).
• ai can be a black box (if you don’t know how it works, and make it challenging to explain the outcomes to your clients).
• like smartphones, you may end up using only a part of ai capabilities (thereby leaving money on the table).
• ai can be an attractive opportunity for your staff (who may leave to join firms that provide such an opportunity).
• ai can be your a(i)ssistant (if you let it be).

and so on …

don’t just use chatgpt and other similar tools for entertainment or novelty. think about how you will know/learn about the ai components embedded into your firm’s tech stack. imagine what your clients can believe and be able to do with more ai in the usual accounting profession-related software solutions they use. invest your time to learn about ai and its capabilities.

posted on march 5, 2023