ChatGPT was released to the public in late 2022, and since that time, ChatGPT and artificial intelligence (AI) in general have become some of the biggest topics in the news and in the business world. Businesses have been analyzing how they can best use AI to improve their processes. Policy makers have been examining which regulations they should apply to the use of AI for the protection of the general public.

In 2024 the European Union passed its Artificial Intelligence Act, which was the first comprehensive regulation on AI by a major regulator anywhere. Prior to that Act, AI regulation was haphazard with several US states and municipalities having passed limited regulations regarding AI.

But what is the impact of these regulations? Gies accounting faculty William Ciconte, Andrea M. Rozario, and Oktay Urcan recently completed studies on the impact that AI regulation has had on the business sector. Their results indicate that AI regulation can have both positive and negative impacts on business.

Corporate risk

In their paper titled “Artificial Intelligence Regulation and Investor Risk: Evidence from State and Local Artificial Intelligence Mandates” they examined the impact that AI has had on corporate risk. Overall, they found that AI regulation has had a positive impact on corporate risk.

Their study showed that laws to curb AI-related misuse are perceived favorably by companies’ shareholders, and the laws bring about real actions by firms to comply with these laws. “So we wanted to examine whether AI is reducing firm risk, and we find evidence for this,” said Urcan, a professor of accountancy and academic director of iDegrees at Gies.

The regulations encourage firms to hire executives whose position entails monitoring potential harm from AI and ensure compliance with the regulations. The firm thereby reduces any fines or other sanctions that the company could otherwise face. And investors are reassured by this and see these activities as beneficial. Essentially, AI regulations are reducing firms’ risk because the firms are being proactive to minimize that risk.

Impact on Innovation

AI regulation has had a negative impact on innovation, as they examined in their white paper “Do AI Laws Inhibit Innovation?” Their research indicates that AI regulation can negatively impact innovation. This negative impact results in large part from the inconsistency and uncertainty about the regulations.

Not every state or municipality currently has laws and regulations related to AI, and those that do are focusing on different misuses of AI. “The laws are all very piecemeal,” said Ciconte, an assistant professor of accountancy at Gies. “At the federal level in the US, there are still no actual laws.”

The researchers point out in their paper that innovation can have a high risk of failure. The piecemeal nature of current AI regulations in the US causes confusion as to what may be permitted. Firms then become hesitant to engage in innovative activities.

For the researchers, it is important to understand what unintended costs AI regulation and regulatory uncertainty may have on innovation.

Balancing positive and negative impact

The two studies demonstrate that there can be positive and negative results from AI regulations.

“We essentially studied the benefits and the costs associated with AI regulation,” said Gies Assistant Professor of Accountancy Andrea Rozario, “and we believe that our findings are important to the global discussion on how or whether AI should be regulated.”

The laws are fairly new, and they are not universally in place. “There's just been some executive orders, and there's the very recent EU act,” said Ciconte. “So, while the laws may not be the most comprehensive laws, it's useful to get some preliminary evidence, especially since everyone's considering, what's the next step?”

Urcan explained, “The problem here is that most of this regulation, both in the US and around the world, is quite piecemeal, quite all over the place. Current AI regulation has limited scope, but creates regulatory uncertainty.”

The solution, according to the researchers, is to carefully examine the impacts that these regulations can have. “Regulate, but go with a lot of empirical evidence,” said Urcan, “which is what we are showing in this work. This is, I think, a very fragile type of a regulation, meaning that we need AI regulation, but we shouldn't overdo it. There is a high risk that over-regulation can prevent AI from realizing its innovation potential.”