The Role Of AI Is To Scale Human Attention, Human Judgment, and Human Agency

Chris Shelton, Senior Vice President, Chief Product Officer and President, AES Next, of the energy company AES Corporation, shares his insights in a conversation led by Adam Bryant and David Reimer. The discussion highlights the importance of adopting AI through clear constraints, integrating AI into transformation processes, and leveraging AI for innovation. Shelton emphasizes cultural initiatives, the role of transformation offices, and embedding AI in strategic thinking as key components.

Reimer: What is your framework for thinking about how to best adopt and integrate AI into your company?

Shelton:We have three pillars for how we think about this. The first is based on being clear about constraints—what’s not good enough and what could be better. The constraint related to AI is the adoption rate. The organizations that have the best adoption rate are the ones that should be expected to thrive, and the ones that have the worst are likely not to do well. So absorbing at the fastest rate became a cultural initiative for us.

The second pillar is our focus on transformation of the way we work. We have a standing Transformation Office in the company called Energy Star, and that helps us focus our work around AI. The big imperatives around AI are managed there. We don’t need a separate AI program because we already have a transformation program that integrates all the business problems and stakeholders with the technology. We don’t have to reinvent all of that.

The third pillar is a focus on new offerings and making sure that we’re not using this earth-shattering technology for incremental innovation to make our business 10 percent better. We need to make sure we’re not missing a broader tectonic shift that the technology enables.

Bryant: How have you encouraged adoption in a meaningful way?

Shelton:We start by telling everyone that legal is not going to have a problem with them using these tools. So step one is telling people that it’s not going to hurt you. It might be better than you think. It also might be worse than you think, but the company still wants you to use it.

The practical test for us, ultimately, is that when people come to, say, a long-term forecast planning meeting, they have embedded AI in what they’re planning. That’s the question—are they going to change the long-term forecast of the company or the bottom line of the company because of what they’re seeing with AI? Are they willing to make the commitment? We are always asking ourselves how we can implement AI over the next few years to make sure we deliver on our imperatives. And those imperatives are not just around dollars; it’s could also be about using AI to improve safety.

Ultimately, it goes back to the culture. If people don’t want to use AI, what’s the reason? What assumptions are they making about the technology? Are they making pie-in-the-sky assumptions that aren’t grounded in the reality of what the technology can do, and fearing that it will fail? We see that.

Some people are saying that they are not going to depend on AI to give them some deep insight about a particular issue. But then we tell them that we are not asking AI for deep insights. We’re asking it to scale our attention to information, or to filter our attention to information, or to gather information.

We stay focused on the things that we know AI can do. For any organization that prides itself on running a tight ship and being lean, there are of course going to be constraints. So AI relieves those constraints in a lot of ways. If people are business-minded, they will recognize that they can use this tool to do things that they might otherwise not be able to do.

Reimer: Any caution flags you would raise for leaders of other organizations who are implementing AI?

Shelton:I mentioned this a minute ago, but a big caution flag is expecting insights from AI. Might that happen someday? Yes, but we don’t want to set expectations in ways that are going to lead to people being disappointed, which could then set everything back.

I do think AI can provide insights, but I don’t think people are going to bet on it for meaningful, high-stakes decisions. You don’t want to set the expectations too high, but it doesn’t mean that the outcome can’t be amazing.

Another way to look at this is that many people are worried about losing their jobs to AI. But humans have judgment, expertise, and agency. Those are unique things, especially in combination.

So what’s the benefit of the AI? As I said, it scales human attention, it scales human judgment, and it scales human agency in ways that humans alone can’t do themselves. Every company should be challenging themselves to think about how they can be embedding AI into their strategy. Think about a flour mill in 1900. The people running it at the time would have been making a mistake not to think about how they could use electricity.

Andrew Ng, AI visionary and AES partner in the AI Fund, has famously compared AI to electricity in terms of its impact. One reason this is a great comparison is that electricity is like raw work, and raw work can be spread over everything. AI is raw intelligence and it too can be spread across everything. So organizations should be thinking about how to leverage it as part of their strategy.

Bryant: It’s clearly a meaningful challenge for you to wrestle with all the unknowns around this new technology. Where does that come from?

Shelton:I think most people, and this is certainly true for me, want to make things better. They want to see people flourishing. Electricity is a great example. There’s a direct correlation with electricity adoption and human flourishing. It goes hand in hand. There’s an opportunity for artificial intelligence to be that way.

I work in the electricity industry, and we take pride in working on what we call societal infrastructure. The idea of intelligence spreading is very similar to electricity, so it’s very exciting.