NEW YORK - Business executives are more engaged in investigating potential uses of generative artificial intelligence tools for their organizations than they were a year ago, according to a recent survey by the AICPA & CIMA. But actual implementation of the transformative technology in key operations remains rare.

At the end of last year, a majority (56%) of business executives said they weren't yet considering the integration of generative AI tools within their businesses, according to the quarterly AICPA & CIMA Economic Outlook Survey. This quarter, the percentage of business executives not yet contemplating the use of generative AI dropped to 38%.

Still, while 30% said they were experimenting with generative AI in business applications, up from only 23% last year, the percentage of companies that are fully using generative AI in key operations hasn't budged much. Last year, it stood at 4%. This year, it's 6%. Leading categories for usage are core business operations, finance and strategy and sales and marketing.

"The data suggests we're right on the cusp of broad adoption of Gen AI," said Tom Hood, AICPA & CIMA's executive vice president for business engagement and growth. "We're rapidly moving from the contemplation to experimentation phase. This technology has seen faster take-up rates than other adoption cycles, so we expect implementation to move quickly as well."

One sticking point for executives looking to take the plunge on generative AI is its perceived risks. Those appear to be moderating as business leaders become more familiar with the technology, but concerns still remain high.

Some 92% of survey respondents said they had at least a slight concern about the potential for privacy, ethical and accuracy risks associated with the use of generative AI tools, with a little more than a third (34%) saying they had significant concerns. The latter figure is down from 40% last year, however.

Improvements are needed in data security for safe, responsible and ethical AI usage. While 35% of business executives said they had security policies and protocols in place, 58% said they did not.