For accountants, what is AI, anyway?

The sheer number of times people have written, spoken or even thought the phrase “artificial intelligence” has spiked dramatically in just one year since ChatGPT — possibly its most prominent specimen — exploded onto the scene. 

But just what is it we’re talking about when we discuss AI? Amid the panoply of sales pitches and marketing claims that have sprung up like mushrooms after a hard rain, it’s easy to lose sight of the basics. 

If we can’t establish some common baselines about what AI is and is not, as well as what AI can and can’t do, it will be highly difficult to have the coherent conversations on the topic that this profession needs. 

This is why, as part of Accounting Today’s inaugural AI Thought Leader Survey, we wanted to ask some very basic questions about AI and its capabilities — we asked 22 AI experts on AI and its future in accounting not about the bleeding edge of this technology and its application to the accounting world, but about these very fundamental and basic concepts that are practically prerequisite for answering the larger questions that have captured people’s attention. 

Looking first at just what people think AI actually is, we see a variety of different takes. Some respondents, like Automata founder Wesley Hartman, set their bar very high for what counts as AI in a way that excludes what many today understand the term to mean. 

“Artificial general intelligence is my definition of AI. This is where a system will use reasoning, planning, learning, natural communication, and the ability to adjust when new information is presented to accomplish a task,” he said. “I also think that the creation of new tasks or goals are important to distinguish between what is AGI versus what is close to AI versus human intelligence.”

Others, like PwC vice chair Wes Bricker, took a much more expansive view to include anything that enhances our capabilities as humans. 

“Artificial intelligence is advanced technology designed to enhance our capabilities as humans. It allows us to put people and technology together to create something bigger and better. It is a type of deep learning that uses prompts or existing data to create new content, including text, code, images, videos and audio,” said Bricker.  

The majority staked out ground somewhere in the middle, with a wide variety of personal definitions, but what they tended to have in common was that AI is something that can perform tasks that usually require human thought. Jin Chang, CEO of advisory services platform Fieldguide, elaborated a little further on what he thinks that means. 

“Generally AI refers to the simulation of human intelligence, where machines can perform tasks that typically require human intelligence, such as problem-solving, decision-making, and understanding natural language,” he said. 

In the first of what will be a multi-part series, we explore further how AI thought leaders respond to the question “What is your personal definition of AI? What is artificial intelligence?”

Tune in tomorrow for what these same people think does not count as AI.