It's impossible to talk about workforce innovation without addressing artificial intelligence.

During a virtual roundtable, we asked Business Insider's Workforce Innovation board to tell us how AI would transform companies over the next year.

One major topic was reskilling and upskilling employees. BI's Tim Paradis, who focuses on the future of work, asked whether there were any fears that dire predictions of AI outpacing human capabilities to learn, like those advanced in a recent article by Sun Microsystems cofounder Vinod Khosla, would prove true.

"The basic drivers of job creation will change with a technology that does not only augment human capabilities, but also may surpass them altogether," Khosla wrote, "leaving education and upskilling — a historical avenue for retraining and relevance — somewhat impotent."

Board members rejected the grim prognostication.

"I disagree with that statement," said Justina Nixon-Saintil, IBM's vice president and chief impact officer, though she agreed that AI would exacerbate the digital divide.

"We made a commitment to scale 2 million people in AI by 2026," Nixon-Saintil said. "There's a sense of urgency in making sure we are not leaving people behind."

AARP's Marjorie Powell argued that the global labor shortage was evidence of an enduring need for human workers but that no one can afford to be complacent. "You're going to have to relearn how to do your job," Powell said.

Others described the importance of communicating clearly with workers about what AI means for their jobs now and in the future.

"We're trying to figure out how this transformation is going to occur," said Chris Deri, the president of the corporate advisory businesses at Weber Shandwick. "There has to be transparent communications and stakeholder engagement alongside figuring out these enterprise and strategic matters."

The following has been edited for length and clarity.

How will the most innovative companies be transformed by AI over the next 12 months? Justina Nixon-Saintil, vice president and chief impact officer, IBM "Any tasks or jobs in the company that could be automated by AI will happen within the next year."

A lot is going to change. Any tasks or jobs in the company that could be automated by AI will happen within the next year. For any jobs that cannot be automated, workers will need to understand how to work alongside AI.

You need to understand what AI means for your role, how to use it to increase productivity or efficiency, and how to create new products and solutions leveraging AI. That's going to be the big shift, and here we actually continue to upskill our employees to prepare for that shift.

Last year, we had an AI challenge across the company, and a huge percentage of our employees participated. They had to get training first, and then they were part of a team to come up with innovative solutions leveraging AI. We did it again this summer. So the company is very focused on how to upskill all employees, because it's going to impact every role in the company.

Anant Adya, EVP, service offering head, and head of Americas delivery, Infosys "How can we eat our own dog food right before we start telling customers to adopt AI?"

We have two objectives. One is to become an AI-first enterprise. The first task is to see how we can start using AI in everything we do, whether it's in our claims or travel expenses, procurement, or handling invoices.

Our philosophy is very simple. AI is not about replacing an individual or replacing an employee but how we can amplify our productivity by using the tools. How can we eat our own dog food right before we start telling customers to adopt AI?

Second is that for all the offerings that we have today, we want to make sure we infuse AI into them. When we go and talk to our customers about task-level automation, coding or developing an application, or migrating a workload into the cloud — how can we infuse AI into everything that we do? So it doesn't just become a people-based offering but a people-plus-AI-based offering.

Marjorie Powell, chief HR officer and senior vice president, AARP "It's important that everyone learns how to deal with prompt engineering; we are not going to hire prompt engineers."

We believe there's going to be a greater emphasis on upskilling and reskilling alongside AI technologies.

It's important that everyone learns how to deal with prompt engineering; we are not going to hire prompt engineers. It's going to become a skill for employees to craft clear and specific prompts, adjust to the right audience, and utilize elements and problem-solving skills to identify inaccurate responses.

It's about adaptability. Employees are going to have to be open to continuous learning and adapting to new AI tools and technologies in their jobs. So we make sure we're providing some kind of atmosphere and culture for being excited and energetic about learning.

The workforce will have to problem-solve regardless of the tool and the technology, and it will require test-and-learn approaches. We created a "community of practice" in the company where individuals can join and help with peer learning and have a sandbox to play around with AI, depending on where they are.

They can cross-pollinate across business-unit lines, and that has become very exciting. We also use that group to pilot the new technologies we're considering rolling out to the company.

Data literacy is going to be very important — understanding how to read and interpret and tell a story of the data.

We're also developing power users in the company, people who are really, really good at this that other employees can tap into when they need help.

Alicia Pittman, global people team chair, Boston Consulting Group "Gen AI also offers an opportunity to avoid leakage of human capital, which obviously we all care quite a lot about."

We have 35,000 employees, and north of 70% are active users of generative AI now. We've taken a full-company approach of upskilling and rolling out both custom and off-the-shelf tools, regardless of the job.

The bottom-up innovation power of gen AI is super exciting. For example, our teams and employees have created more than a thousand custom GPTs, mini automation machines, to-do scripts, and dashboards to create bespoke analyses and build newsletters — it's super quick, and it doesn't require big investments or processes.

That's one thing that's specific to gen AI versus other types of technology innovation — it's sparking a lot of energy and unlocking human capacity.

Gen AI also offers an opportunity to avoid leakage of human capital, which obviously we all care quite a lot about. It really changes the game in terms of the amount of learning you can keep within the team and not lose as people move around.

We're also seeing how gen AI is creating an exoskeleton for generalists to be able to take on more technical tasks. It's not going to turn a generalist analyst into an expert coder, but actually our analysis would show that they can get 80%-plus there on some coding tasks, which then you still need the expert coder to do quality control.

Maggie Hulce, chief revenue officer, Indeed "Do I have to build this huge team of AI engineers? The answer is no."

We're trying to infuse AI in all our externally facing products and solutions to make them more insightful, intuitive, and delightful to use — and, in many ways, to create experiences that mimic what a personalized human experience might be like, but at scale.

We also see most roles and workflows becoming augmented with AI. That's becoming more and more obvious as the months go on and as teams engage with AI and find applications and new ways of working.

I was at Dreamforce recently in San Francisco. What Salesforce unveiled was how it wants to help companies essentially launch AI agents in sales, customer service, marketing, and commerce.

For many companies to get the full power of what AI can do, you have to be able to link up lots of different datasets that are usually in disparate places. It's hard work, and it's very time-consuming. To hear the stories of companies being like, "Guess what, you don't actually have to move that data around anymore," this is going to change how a lot of companies experiment and start to play with AI-oriented applications that will fundamentally change their operations.

There are a lot of companies out there that have already started to light it up. We're thinking a lot about how sales, customer service, marketing, and product have operated for a long time in various silos. When you think about customer experiences, they really shouldn't be that way. They should be much more integrated than they are today. Some of these technologies can be part of the accelerants for how customer experiences can be reimagined.

Part of what's been hard for people is like, "OK, but do I have to get all of these disparate parts of my company to all work together in a new way? Do I have to build this huge team of AI engineers?" The answer is no. You don't have to be the tip of the spear in terms of sophistication.

Kenon Chen, executive vice president of strategy and growth, Clear Capital "As AI accelerates, you also have to accelerate the handling of data in a way that can produce accurate models."

At Clear Capital, we participate in the highly regulated housing-finance industry. To enable AI to be used in accordance with our clients' expectations of governance and quality control, we've taken a three-pronged approach.

First, we've been investing in creating good industry standards and working with the industry to ensure there's agreement on how AI models should be tested, how they ensure compliance with antidiscrimination and fairness practices, and how the data should be handled.

Second, we've been investing a lot in higher-quality data. As AI accelerates, you also have to accelerate the handling of data in a way that can produce accurate models. So we've invested quite a bit in better digitization and generation of data.

Third is fostering collaboration between innovators, industry players, policymakers, and regulators. We participated this year in the Federal Housing Finance Agency's generative-AI tech sprint, where about 80 leaders from different aspects of the housing industry brainstormed ways to apply generative AI to the lending process to improve borrowers' experience.

One thing that became clear after participating in that was where we have more tools and more capabilities to accelerate. We always have to go back to the why: What is the mission? And we have to make the mission at the company level that much more clear so that as we apply these new tools, it's always for the sake of the mission.

Chris Deri, president, corporate advisory businesses, The Weber Shandwick Collective "There's a mega-trend need for all of us to adopt AI and understand it, but that means something different for different parts of our organization."

We're a public-relations firm, so we're looking at AI for ourselves and for our clients.

In terms of our own people, one lesson we've already learned is that there was an instinct to take a top-down approach, as in everybody's got to figure out how AI is going to change their business, but we need to take a bottom-up approach.

In the simplest sense, we do consumer marketing, healthcare marketing, and corporate advisory work. We really need to design training for each of the different practitioner types.

There's a mega-trend need for all of us to adopt AI and understand it, but that means something different for different parts of our organization.

Another thing we're doing is creating a futuristic think-tank group in our company — three or four people who have been putting together an ecosystem to give us constant learning. We also have academic institutions, investors, technologists, and other practitioners to help.

We're also looking at our business model and how we budget and scope. The classic assignment we get is companies call us and say, "We want to map our stakeholders and what their expectations are of us." That used to be a project that we would figure out with just consultant time. Now we're applying AI to do some of that.

We need to figure out how to take advantage of those efficiencies to add new value — that human value. So we need to relook at our entire business and service model, and we haven't figured it all out yet.

Anant talked about "eating your own dog food." We've mapped out the implementation of that stack for ourselves to test and learn, and then we will start to use that in how we show up as advisors to our clients on the same journey.

Sharawn Tipton, chief people and culture officer, LiveRamp "What we're trying to do is really keep people at the center of this transformation."

We're thinking a lot about the AI trust gap. I'm in San Francisco today, and I was walking around and I saw all these autonomous vehicles, and it dawned on me that my mother would never ride in one. She just does not trust that technology.

As a data company, we're thinking a lot about how do we ensure that our customers, our clients, and our partners trust that we're going to protect their data and use AI when appropriate and be certain that we have safeguards for when it isn't appropriate. And then I think the biggest area that my team is focused on with this trust gap is job security and what this means for the average worker.

So what we're trying to do is really keep people at the center of this transformation. We're thinking a lot about employee resource groups. That way we've been able to reach all different types of team members. So how do we target our ERGs and think about who are the slow adapters? Who are the folks that typically get left behind, and how do we give them a little extra push?

What we want to do is think about how to continue to have our team members trust their leaders, but also trust AI and where we're going, and get them to understand this isn't about replacing jobs. We are really focused on transforming jobs. That is a huge shift and a huge opportunity for us.

Shane Koller, senior vice president and chief people officer, Ancestry "The more people we get thinking about governance, the better AI technology will be used."

There are a lot of different points of view on what AI will and will not do and how quickly.

I was reading an article in The Wall Street Journal over the weekend that said for every one person who believes that AI is going to replace everyone's jobs, there are three or four who think there's going to be this boom in AI and then it's going to settle down. And that the promise we think AI is going to deliver will probably take longer than we think to come to fruition.

One of the things that we've been doing at Ancestry very intentionally is bringing up the subject of governance, because I think we all think of AI as this steamroller that's going to come and either we're going to step aside from it or we're going to get steamrolled.

The more people we get thinking about governance, the better AI technology will be used. And I say that within Ancestry as well as within companies in general and within governments. We have more of a role to play in how this rolls out than we think.

We've been building AI into our product for over 10 years now, and we've been giving more thought to how would we govern the use of AI within our own company so that we're not, for example, putting our customers' data at risk. We're very conscious that security and privacy is first and foremost in every discussion. It's really about how you govern such an amazing technology.