Enterprises Say AI Regulation Won’t Stop Corporate Innovation

With Democrats likely to focus on reining in AI, business tech leaders aren’t worried that new policies will affect their technology plans.

Originally published on www.wsj.com

WSJ

Corporate technology leaders say they are pushing ahead with their artificial intelligence plans—regardless of whether AI regulation solidifies under Democratic control of the House and Senate and state offices.

Regulation around AI models could take any number of forms, especially as Democrats are likely to focus on the topic of safety, which had already become a bigger priority for lawmakers. Democrats at the national and state level may push for regulations that set guardrails for new AI models, which could include testing models before they are released publicly.

Prominent companies and individuals have also recently called for mechanisms that would allow AI labs and countries to slow development in a coordinated way to buy time for alignment research. Lawmakers from both parties have proposed many bills to address rogue AI systems.

Tech business leaders who spoke to The Wall Street Journal Leadership Institute said none of those scenarios is much of a concern for their own AI initiatives, which have only been gaining steam as they increase spending on AI.

There are a variety of reasons: Some tech leaders say they have been pursuing their own internal guardrails and standards for AI safety, making regulation less likely to affect how they use AI. Many also believe the existing capabilities of AI models are enough to satisfy their technology needs, making a potential AI development slowdown unlikely to alter their future AI plans.

Prakash Kota, chief transformation officer of UKG, said AI regulation could give companies like the HR software firm the comfort to use AI more aggressively.

“Right now it feels like there’s unknown territory that I’m going into,” he said. “It’s a lot more on enterprises to define those things. If policies come into play that give us more comfort, risk-averse people would want to lean in.”

Dave Treat, chief technology officer of Pearson, said the London-listed education company had already been focused on using AI responsibly within the firm, and that won’t change as the regulatory environment evolves.

“We already operate with strong governance, privacy, security and human oversight standards, and we expect those principles to remain important under any regulatory framework,” he said. “Thoughtful regulation could help build trust and provide clarity.”

‘There’s a long way to go’

At the same time, Treat said existing AI models offer plenty of untapped capabilities. So even if the pace of frontier AI model development were to be slowed, Pearson would still find current AI models capable enough for its needs.

“With the power of the models as they are today, there’s a long way to go,” he said. “We’re just at the beginning of diffusing or achieving all of that value, so it wouldn’t be the end of the world. But obviously I’m a big fan of constant innovation.”

Doug Gilbert, chief information officer and chief digital officer of IT and business services company Sutherland, said the need for the most powerful models from OpenAI and Anthropic is “extremely low” for everyday applications. “If you look at 95% of the tasks that need to be done, you can use models that have been around for a while,” he said.

And though it’s possible that pausing the development of AI models could hold back innovation at customer-experience tech provider Genesys, Chief Executive and Chairman Tony Bates said the capabilities of current frontier models, and its own models, are good enough.

“The technology itself is less of the discussion,” he said. “The conversation is moving from simply pushing the frontier to making sure AI can operate safely in the enterprise.”

From models to harnesses

Some enterprises are also shifting focus away from the AI models themselves, and putting more emphasis on the model harness instead.

The harness is software that actually runs the AI system and does things like provide memory and business data context for the model. It also allows businesses to mix and match different models, and manage AI token costs—ultimately making enterprises less reliant on a single AI model provider.

“Regulation in terms of slowing the pace of model innovation, would that hurt us on some level in terms of an enterprise? Sure,” said Alex Balazs, chief technology officer of financial-software maker Intuit. “But we right now are making a huge focus and investment on building the right AI harness.”

As the regulatory climate changes, he added, Intuit is prepared because it had already been putting its own responsible AI initiatives into practice. “Our principles go beyond what the law requires and reflect what we believe should be the industry standard,” he said.

Staying the course

Other tech leaders say they are similarly prepared for any AI regulatory environment: An enterprise’s own controls and governance over the technology take precedence.

“We built for it. Every agent we run has an audit trail, a confidence threshold and a human checkpoint,” Sutherland’s Gilbert said. “The companies at risk aren’t the ones using AI aggressively, they’re the ones who can’t produce a log of what their AI did and why.”

Andrea Schulze Dias, vice president and group chief information officer of Toshiba America, said that regardless of the regulatory timeline, responsible AI governance remains a priority.

“While new regulations may introduce additional requirements around data privacy, security and governance, our focus remains on adopting AI responsibly,” she said.

State-level policy ahead

As Democrats begin to hone in on AI regulation, especially for data centers, many corporate tech chiefs said they expect AI policy to remain at the state level, rather than moving quickly federally. And, they expect any change to take time.

“The midterms settled the one question that mattered for many enterprise technology leaders: There will be no federal ceiling on AI regulation anytime soon,” Gilbert said. “State laws—California, Colorado, New York, Texas—are the operating standard, not a placeholder.”

“I don’t expect sweeping federal legislation from a divided government, but I do expect aggressive oversight,” he added.

Still, there are downsides to dealing with a patchwork of state-level AI laws.

Matt Calkins, chief executive of AI automation software company Appian, said that state-level regulation could hurt companies because “it could be irregular across the nation, and because some issues, like alignment and intellectual-property rights, are best handled federally.”

And while the outcome of the midterms does increase the likelihood of AI regulation, Genesys’s Bates said the more important issue is “whether we get practical guardrails that build trust without slowing innovation.”

About The Author: Belle Lin is a reporter writing about AI and enterprise technology for the WSJ Leadership Institute’s CIO Journal. Belle’s reporting explores how the latest advancements in AI affect businesses, as well as topics like semiconductors, data centers, cybersecurity, open-source software and tech hiring. She is also interested in the ways new technologies affect traditional industries.