NEW JERSEY — NJBIZ recently hosted a panel discussion about how rapidly artificial intelligence is evolving and affecting society in myriad ways.

The 90-minute discussion, moderated by NJBIZ Editor Jeffrey Kanige, featured a panel of experts comprising:

• Joshua Levy, general counsel at Gibbons PC and director of the Business & Commercial Litigation Group.
• Carl Mazzanti, co-founder and president at eMazzanti Technologies.
• Hrishikesh Pippadipally, chief information officer at Wiss.
• Mike Stubna, director of software engineering and data science at Hackensack Meridian Health.
• Oya Tukel, dean of the Martin Tuchman School of Management at the New Jersey Institute of Technology.

The discussion opened with the panelists discussing how their respective organizations are deploying AI both within their own organizations and for stakeholders. (See video.)

‘Where do I start?’

A lot of companies these days are seeing the growing number of AI tools available and asking where to start.

Hackensack Meridian Health’s Mike Stubna said one of the most important beginning steps is weighing the risks associated with any AI tool.

“There’s a few components of our overall strategy. As I mentioned, one of the big components is a comprehensive risk assessment,” said Stubna. So, before adopting any sort of new AI-powered tool, it’s important to attempt to understand and quantify the risks associated with it.

“For example, maybe you’re using a service to summarize pages and pages of information. If you’re sending this service proprietary information, and they have the ability to save that and use it for their own purposes, that’s a risk, right? Because that might be leaked out by them. So, understanding the risk is a really big part of the strategy for every solution.”

He said that the other component is the value that the solution provides.

“So, in many cases with, especially workflow optimization or administrative efficiencies, the value could be quantified in terms of hours saved,” Stubna explained. “Of course, vendors will tell you that their solution will save so much time, but it’s really important to do your own due diligence.”

AI guardrails

Hrishikesh Pippadipally said that when thinking of an AI strategy, business leaders need to establish responsible AI governance.

“What’s your company policy? What software can we use?” the Wiss CIO asked. He noted that his firm has strict guardrails and approvals on what can and cannot not be used. “So, that’s more from the risk perspective. But moving on from there, in terms of adoption — and how do we structure this within our firm — I would probably phrase this in terms of people process technology.

“Technology is only as good as those using it. If you don’t know how to use it, there’s no use to it. So, we have done a big round of upskilling AI training six months ago.”

He stressed that the training process is important because it entails actually going through real use cases from the pilots that the company conducted.

Joshua Levy said, “I certainly agree that the evaluation of the risk is paramount. And I think I mentioned before, for us, confidentiality of our data and our clients data is front of mind — and then also scrutinizing the value. And so maybe to approach it from an angle that hasn’t been discussed: expense.

“At Gibbons, we are about a 150 to 160 attorney law firm, and we have the overhead we have. But there are firms larger than us that can afford maybe multiple highly overlapping platforms. And I’m sure there are firms smaller that have to make even more careful decisions about which technologies to invest in. It’s an ever-changing landscape, and we try and approach it with humility and understanding that what we may be investing in now might not necessarily be what we need in the future.”

Securing data

Mazzanti said that there is so much “tool creep” taking place because everyone’s offering something and each saying that they are the best. He noted that his firm is seeing trends within clients’ organizations where a heavy number of users are already utilizing a particular tool.

“Maybe we should consider embracing that to the organization because your staff already voted,” Mazzanti said about what he will tell these clients. “They’ve already said that this is the one they like the most and you already have a bunch of heavy users. Maybe we could do that.”

In referencing risk assessment and governance, Mazzanti suggested digging a moat around data.

“So, robotic process automation has been around for a long time. We’ve been feeding it data to have it go do these individual tasks. That was before there was some intelligence or the generative concept came across. Well, when you were feeding the data, it was typically your own set in in your own servers in your own environment being run. And now that it’s generally available and you can rent, not a lot of the tools are offering to put security around your data,” he explained.

“I’m very surprised at customers who deploy without any sort of plan whatsoever,” he said, pointing to incidents that can occur with data being leaked or breached because the proper controls were not set up. “Privacy tags — super important. Start doing that around your HR, your offer letters, your salaries, things like that.”

Staying vigilant

The panelists also discussed the need for training when it comes to using AI tools and how difficult it is to grasp these new concepts.

Tukel noted that the next generation is very tech savvy.

“I have to say, maybe it’s more encouraging for faculty to be in the forefront together with the students using the technology. Because sometimes we are skeptical about what a technology can do,” Tukel explained. “So, we always go back to the fundamentals of what we need to learn. I agree with Carl [Mazzanti] that there are a lot of loose ends with the new technology — up until it solidifies.”

Because of that reality, Tukel said that workers understanding the AI strategy is very critical.

Be transparent

From there, the conversation snaked through a number of topics around AI, such as the limits of what the technology can and cannot do — and should and should not do; the risks and potential pitfalls of the technology; cybersecurity; the workforce impacts; regulatory issues; inclusivity and more.

Tukel stressed the need for AI transparency.

“I think it is OK to not look at this tool as a shortcut that covers our areas of deficiencies. It’s a tool that helps us,” said Tukel. She noted, however, that the technology still has its problems in terms of bias in the data and algorithm. “But declaring that this was prepared by me — but using AI tools — in your documentation for public-facing writings and news you are putting out, can definitely put you in a better position.”

“[J]ust know that if you come to Hackensack Meridian Health, that we’re on the forefront of using AI,” said Stubna. “And you can have a lot of confidence that it’s something that we take extremely seriously — ensuring that this is used in a responsible way that really focuses on patient care first and foremost.”

“At Wiss, we have heavy transparency,” said Pippadipally. “We try to use AI to enrich the product of client outcomes as much as we can. Anything, any financial issues that you guys have, feel free to reach out.”

“If we had to leave someone with parting words here – it would be, choose a good partner that’s done this – binds your vision and values and know you support to walk hand-in-hand with your suppliers,” said Mazzanti, stressing service delivery. “My organization is incredibly partner friendly. We work with your team to evolve with your best interest at heart. You’ve heard from some great panelists here today.

Stay flexible

“My own approach to this is, simply, keeping my eyes open and trying to stay humble in this ever-changing environment,” Levy said. Even if I’m here, even if all of us are here, because in some ways we’re experts in this arena, I don’t know that I think anyone truly has a perfect understanding of where any of our industries are going to be in five years, 10 years, 20 years.

“And we just have to be flexible and keep that in mind — and work with the right folks to navigate the future.”