Although artificial intelligence has the potential to transform a wide variety of industries, including manufacturing, both government and industry need the right policies in place to responsibly use and promote trust in the technology, speakers said at the SelectUSA Investment Summit.
Enacting policies that support AI-related workforce training and K-12 education initiatives, along with creating a more cohesive regulatory environment across states and shoring up the power grid, topped the wishlists of industry experts speaking at the U.S. Department of Commerce event in National Harbor, Maryland.
Olivia Negus, director for trust, data and technology policy at the Information Technology Industry Council, said she is excited about the way AI has made a variety of industries more productive.
“We really are seeing transformational applications across all sorts of different sectors of the economy,” she said during a May 4 session. “I think we’re just at the beginning of that.”
At the same time, Negus said both the federal and state governments need to develop a cohesive regulatory environment that promotes greater certainty and innovation.
Negus cited the Trump administration’s 2025 AI Action Plan, which sets several broad goals, including creating high-paying jobs as the U.S. builds out AI infrastructure, building AI systems that are “free from ideological bias and designed to pursue objective truth,” ensuring data security and mitigating “unforeseen risks.”
She also pointed to a December 2025 executive order that called on states to promote innovation by eliminating “excessive regulation” around AI. For example, the order cited a “patchwork” of state-by-state AI regulations that industry can find challenging to navigate. It also called out state laws that require entities to “embed ideological bias” within large language models or attempt to regulate AI beyond state borders.
The executive order called for a national policy framework, established a litigation task force to challenge certain state laws and also restricted certain funding for states with such laws.
For AI to fulfill its potential, companies need to set flexible policies while establishing a secure environment that encourages trust in the technology, said Miranda Lutz, senior manager for global public policy at Cisco.
“We’re trying to make sure our customers have the choice they want in terms of AI infrastructure,” which means in part allowing data to be processed locally while promoting an open internet, she said.
At the same time, Lutz said, “security is absolutely foundational for AI,” because public adoption will stall without trust. As a result, she said Cisco is “baking in” security to its networks and all of its AI applications. “We’d like to see the [Trump] administration and Congress recognize the need for security as a critical foundation.”
Lutz said AI security also extends to supply chains, which more and more companies are trying to diversify. For example, Cisco is investing in building its own Silicon One chips rather than relying on other companies.
She added it’s also important to think about AI security in both the physical AI supply chain and large language models, which require high-quality training and secure databases.
“All of that is going to be incredibly important” for AI to fulfill its potential, said Lutz.
Another crucial element of AI infrastructure is power, said Erica Fitzgerald, director of federal government relations at Schneider Electric.
Fitzgerald said companies like hers are “taking a fresh look at power delivery” as they try to keep up with demand and contend with data centers that sometimes overwhelm power grids.
Power “is one of the main things our customers come to us with,” when they want to build a data center or other facility, she said. “There’s a lot of power demand around the entire AI ecosystem in the US.”
Fitzgerald said states that want to attract AI-related facilities need to “show how they’re positioning themselves to power AI.” Without demonstrating that ability, she said, it will be extremely difficult to attract investment.
Both Lutz and Negus said any company that wants to start an AI-related project needs to actively engage with the local community from the outset and show how everyone — not just the company itself — will benefit from the project.
In part, Lutz said, this requires emphasizing that it will produce long-term jobs in addition to temporary construction jobs.
“Communities want to see that investments are translating into benefits for them,” said Negus. She added that companies also should work with local community colleges to ensure workers have in-demand skills and should support workforce initiatives to help those whose careers are in transition due to AI.
Technological change “can be really scary,” she said, and companies would do well to help people navigate those changes.
