You may have heard, in the past, the age-old slogan: The only thing certain is change.

But that’s even more true now, as we ponder the awe-inspiring power of LLMs, neural nets and components of AI technologies. In some ways, this isn’t even just “technology” anymore. It’s a vast power, one with the reach to impact our world.

With that in mind, how do you, as a person, adapt to that change? How does a company adapt to that change?

There’s the idea that people, on an individual level, have to get on the bandwagon, or be left behind in the dust. Come to think of it, that’s true for companies, too. Experts talk about AI as “assistive,” or “human-first,” assuming that the tools will not simply replace human workers.

In that context, I heard a good discussion on this at the April 9–10 meeting at MIT that my organization, Imagination in Action, puts on there every spring.

In a segment at IIA, Jim Rowan of Deloitte interviewed Robert Blumofe of Akamai, Denise Lee of Cisco, Sears Merritt of MassMutual, and Rick McDonald of The Clorox Company about these types of issues.

“What changed first,” McDonald said, in remembering the mad rush to AI over the last few years, “was the amount of data that was available instantaneously versus periodically. So you went from periodic decision making to continuous decision making. And what we found in some of our early digital transformation initiatives is that we didn't do a good enough job of change management.”

Merritt agreed with the importance of handling change.

“I heard a statistic that said that the most successful technology projects have equal change management investments as they do tech investments,” he said. “So change management certainly, certainly a big one. You need good telemetry and observational systems, so that you can govern those tools. We're a regulated entity in the financial services industry, so auditability, traceability, observability, are really important tools for us to have in place as we scale these systems.”

Addressing the colossal reach of new technologies, Lee referenced a reality of “AI first, AI everywhere,” using a bowling metaphor to illustrate how safe AI is appealing.

“You know, who doesn't love bowling with bumper guards?” she said, while conceding that in some ways, the company has had to use “unnatural gymnastics” to scale globally. “It's great. You're going to hit every time. And that's effectively how we've been able to roll out so many tools and systems so quickly, having all the right tools at our disposal, with our own kind of internal systems that we're using, and then function-built for each group.”

Blumofe contrasted “three epochs” of the AI age as successive waves.

“For a lot of the companies that I talked to, there's a similar sort of ‘three epochs,’ there’s kind of a pre-AI epic, which goes up until about the early 2010s, where unless you were a robotics company, or really, an AI company, you probably weren't doing all that much with AI. Everything changed, really, in the early 2010s, with the advent of deep learning, and suddenly you had this tool that could do a lot of really interesting things.”

And then there was the third epoch.

“November of 2022, and you know, that's the beginning of the AI ‘Cambrian explosion,’ where you move from a regime where, you know, AI occupies a few niches, to really an AI-everywhere world,” he said. “You had that ‘aha’ moment, where you realize you're in an AI-everywhere era, and then the question is, well, what do you do about it?”

Blumofe also suggested businesses should build in AI, because otherwise, employees will go rogue with shadow AI.

“We find really, sort of in a grassroots way, ideas coming from all over the place,” he said. “You're much better off embracing these things, and providing the tools, the capabilities, sandboxes, so that people can experiment with this stuff in a safe way, and they don't go off and do the shadow versions. It's going to happen no matter what.”

“You are looking for something where they can go in a sandbox,” he said, describing ideal scenarios. “It's regulated, it's governed. They can experiment and find different ways to leverage a tool.”

Rowan asked Merritt about cybersecurity.

“Cyber risk is really, really on everybody's mind,” Merritt said. “So as we've started to think about our readiness and what we're going to do, it really is a return to first principles.”

Lee discussed energy, in the context of today’s AI market.

“It's the hard ceiling, and it's the foundation, again, depending on who you're talking to,” she said. “And we have taken for granted the grid and electricity for nearly what two centuries, a century, century and a half.”

McDonald urged leader to avoid just relying on tools, suggesting that talent is a crucial part of the equation.

“It really is the leaders’ digital fluency or literacy, their mindset, and the upskilling and re-skilling,” he said. “If those aren't working, then you can buy the greatest tool in the world, and you will not deliver an ROI out of it. So I think that's a huge leadership misconception right now.”

“In a big enterprise, it's still garbage in, garbage out,” Merritt added. “If you don't have your data available in a way that's accessible to these systems, you're still going to have to go back and do that. So I think for companies that have done good jobs on their digital transformation journeys years ago, they're actually in a phenomenal position to take advantage of this stuff at pace.”

Here, we heard a lot about the logistics of setting up the AI systems that are going to drive change in the future. Stay tuned for more from our conferences and the biggest AI news right now.