In early September, a former OpenAI and Anthropic researcher posted to X that AI could kill humanity by the end of the decade. His claim quickly went viral and, according to USA Today, led to Florida U.S. Representative Anna Paulina Luna urging Congress to hold a special session related to AI advancements.
While AI has advanced at an unprecedented pace since ChatGPT entered the technological scene in November 2022, Bryant’s Senior Lecturer of Legal Studies Ronald Washburn, J.D., notes that the federal government’s lack of effort to regulate AI is complicated by the fact that the technology is advancing every day.
“It's possible that the law could keep pace with AI with certain aspects, but the landscape is changing so rapidly that the law needs to be exceptionally elastic,” says Washburn, who teaches the university’s “Legal and Ethical Aspects of Generative Artificial Intelligence” course.
Why AI is so difficult to regulate
The country’s existing laws are ill-equipped to address AI, explains Washburn.
While lawmakers, who are feeling pressure from constituents to address regulation concerns, have been sporadically introducing AI bills to Congress, the proposed legislation has revealed a lack of understanding of the new technology and, in some instances, is simply political posturing.
He notes that earlier this month, United States President Donald Trump announced the creation of an AI task force. Officially known as the Super Intelligence Force, the appointed group seeks to ensure that the nation continues leading within the AI sector and protects American lives and interests.
“We’ll see how that goes,” says Washburn. “I think it isn’t just one office that’s going to do this. It’s going to be several – whether that’s individual task forces or agencies that are created. One agency would be so huge, and it’d be difficult to encapsulate all the different expertise needed.”
A variety of stakeholders need to be at the table for conversations to create effective AI regulation policies, emphasizes Washburn. These individuals include industry leaders, those who are well-versed in the technology, and people with a legal background who also understand the technological side.
What successful AI regulation could look like
Effective AI regulation has to be powerful, notes Washburn.
“It has to do its job and protect while also allowing for growth of the medium, so it can be a greater and more valuable resource,” he says, cautioning against overregulation.
Washburn explains that there's a non-binding, advisory document called the Blueprint for an AI Bill of Rights that the government released in 2022, which details the areas where AI regulation is needed to protect humanity. The document talks about the notion of data privacy and algorithmic bias and, while it has no force or effect on the law, Washburn shares that it’s a starting point for considering policy related to AI; it’s also a document that he has his students refer to when considering how they would form AI regulation laws.
“The main reluctance to regulate comes from the notion that U.S. law only governs in the U.S. (unless other countries agree to accept or adopt). The belief is that regulation, particularly over-regulation, will encourage AI companies to leave the U.S. for countries with little to no regulation,” Washburn says, noting that many industry leaders do not want regulation or would rather self-regulate.
RELATED ARTICLE: Is AI a giant-killer? Bryant expert explains the business revolution
Many AI regulation matters are federally based, which is why an overriding federal law to provide better guidance is needed. But some movement has been made at the state level. For instance, Illinois and Nevada have banned the use of AI to treat mental health, and Utah has placed limits on therapy chatbots. The catalyst for these actions included reports of dangerous responses from virtual assistants, with Washburn noting that any federal AI legislation should determine liability or who has the burden when a negative outcome occurs.
He adds that it’s important to consider political involvement and the role of antitrust laws. For instance, antitrust laws regulating big business often depend on the sitting administration’s interest in enforcing them.
“We want to make sure that we don't have a repeat of that,” Washburn says. “We want to make sure we have strong laws that will be enforced, not based upon who's in office.”
Writing tomorrow’s rules today
Ultimately, regulating AI is extremely difficult because the technology advances so quickly, and the process of changing the law is both difficult and slow.
“We could put laws into place tomorrow, and they could be completely useless in six months,” Washburn says.
It’ll be a lot of work, but this is an incredible time in history, emphasizes Washburn.
“Hundreds of years from now, people are going to be looking back at this great period where this technology became so paramount. It's interesting to be on that ground floor and to understand how these foundational laws will be developed and grow,” he says.