Jeffrey Agnor (Davis, Agnor, Rapaport and Skalny), right, introduces Barton Paulhamus (APL) to a Howard County Chamber discussion on generative AI’s business uses. (TBM / George Berkheimer)

Technology sages have long prophesied that artificial intelligence would disrupt or at least transform business. Nearly 18 months since the launch of ChatGPT, many professionals are still wondering what to do with it.

That much was clear from the questions asked at the Howard County Chamber’s Trending Topics discussion in April, held at the Davis, Agnor, Rapaport and Skalny law firm in Columbia.

Barton Paulhamus, chief of the Intelligent System Center at Johns Hopkins Applied Physics Laboratory in Laurel, presented an overview of the usefulness, pitfalls and current state of generative AI.

Even in his capacity as a principal staff researcher, Paulhamus admitted that he’s found narrow use for the well-hyped technology.

“I use it as an idea generator,” he said. “That’s kind of the state-of-the-art with generative AI right now and what we can expect for it, and where we can really use it powerfully.”

It’s good at compiling information, composing emails, essays and other forms of predictive communication, and performing research swiftly, but it has its foibles. It hallucinates sometimes, or makes things up, a result of biased data, lack of filters for truth and accuracy, and because ChatGPT doesn’t know anything that happened after September 2021.

Nevertheless, Phil Whitebloom, founder of Been There Consulting Services which specializes in sales coaching, said he is an AI believer and uses ChatGPT to do research, compose email campaigns that introduce his services to specific markets, and generate lists of topics that his audiences might be interested in.

“It does a good job of touching a lot of important points and even suggesting some things I wouldn’t have thought of myself,” Whitebloom said. “The downside is that the emails it generates are very similar and can sound insincere. They still need the human touch to personalize them or make them relevant to a target audience.”

Trust but verify

ChatGPT is well known, but there are other models in use.  

“We’re starting to see some curation and there are some repositories of good data, but for the most part it’s [rare],” Paulhamus said.

generative AI tools are changing quickly, he said, and the best advice for businesses and professionals is to focus on how to use whatever tool they’re already using.

“I advise users to verify the information they get and do due diligence before accepting it as accurate,” Paulhamus said.

At present, he said, there’s no good method for ensuring that a generative AI response isn’t copyrighted material or intellectual property. In fact, John Grisham, George R. R. Martin and 15 other authors have sued ChatGPT’s developer OpenAI for using their copyrighted works without permission.

For that reason, Whitebloom advises against using generative AI to write blogs or other content published for public dissemination.

Businesses and professionals should also be mindful that anything they send to ChatGPT goes into the cloud where that information could be stolen, misused or made public.

“There is technology you can host on site so you can use the algorithms without sending anything to the cloud,” Paulhamus said. “It’s expensive, but I predict in another year or two we’re going to see a ton of enterprise services from all the big players. Microsoft already has it.”

Workforce anxiety

Perhaps the biggest fear is that AI could replace workers, but so far it hasn’t happened.

“A lot of studies indicate 40 to 60% of jobs could be disrupted by generative AI,” Paulhamus said. “It could be total displacement, teams of six could be reduced to three, but I think it’s still 20 years away before AI becomes as smart as a human.”

Even then, Whitebloom doesn’t believe it will have dire consequences.

“I’ve been in the tech space since 1980,” he said. “Word processors and answering machines were supposed to replace secretaries, it didn’t happen. They had to learn new skills, but they weren’t eliminated.”

What’s more pressing, said Paulhamus, is the acknowledgment of not knowing where to begin with the technology.

“Some big studies show that about 50% of executives think their workforce is not skilled to take advantage of generative AI, and about 75% of Gen Z feel they are not knowledgeable or equipped to know how to take advantage of it,” he said. “It’s really pointing toward a paradigm shift: how do companies use it and get employees to bake it into their workflow?”

That’s something that is still being ironed out.

APL launched an enterprisewide 10-step process platform called Level Up that introduces staff to generative AI through tutorial videos and asks them to think about how to use it in their day-to-day work.

Although APL’s platform isn’t available for businesses, Paulhamus recommended a new online platform, besuper.ai, that launched on April 10.

“We’re leaning in, generative AI is the future,” he said. “We’ve got to upskill to use it.”