The Future of AI: Experts Discuss Balancing Innovation and Regulation

Benefits & Risks of Artificial Intelligence

The panel examined the importance of managing the emerging regulatory landscape and the need for federal, state, and international coordination on AI policy. They highlighted how addressing potential downsides, such as the misuse of AI for nefarious purposes or the erosion of privacy, is vital and timely.

Andrew Gamino-Cheong, co-founder and CTO of enterprise software company Trustible, said one downside of AI is that it’s being used by people who don’t understand it. “A lot of people are throwing ideas out there, like chatbots, that aren’t safe,” he said, citing Snapchat’s recent chatbot release. “The safety and dangers of how children interact with that are completely unstudied right now,” he cautioned.

AI has plenty of benefits, in addition to risks. There are a number of tasks this technology has the potential to automate, but its ability to quickly analyze and summarize large chunks of text stands out as particularly significant for government affairs. Bill Frischling, vice president of AI at FiscalNote, said he’s excited about the latest AI developments. “What I love about technology is the ability to automate and simplify.”

But AI has many advantages that go beyond the basics of generative AI like ChatGPT. Rep. Donald Beyer, D-Va., who has represented Virginia’s 8th District since 2014, is currently pursuing a master’s degree in machine learning at George Mason University. Beyer is particularly excited about the impact of AI on health after losing one of his best friends to pancreatic cancer last year. Now, it may be possible for AI to predict future pancreatic cancer up to three years before a diagnosis.

AI Policy Areas to Watch

Some of the key policy issues Congress is addressing include data privacy, algorithmic accountability, and the use of AI in high-risk situations such as nuclear weapons launches.

“I feel very strongly that we need to address this at the federal level,” Beyer said of AI. “It’s cool when a state moves forward, but then you just get this crazy patchwork.” However, Beyer pointed out that rarely does a big piece of legislation get it right from the beginning. His hope is that there will be “a whole lot of small, meaningful bills” regulating AI.

One key piece of legislation is the American Data Privacy and Protection Act, which deals with the reality that personal information about American citizens can be collected, finely sorted into meaningful information, and sold.

The Algorithmic Accountability Act of 2022 is another important bill that was introduced by Rep. Yvette Clarke, D-NY. Clarke has asked the Federal Trade Commission to write rules for evaluating highly sensitive systems.

Another bill of note is the Block Nuclear Launch by Autonomous Artificial Intelligence Act of 2023. “If we’re worried about artificial intelligence being a potentially extinctive event, nowhere is that more plausible than with the launch of nuclear weapons,” Beyer said. This bill “says no, AI can be involved in that chain of launching nuclear weapons. It has to be human beings making that decision.”

The development of AI has not escaped the notice of local and state governments either. Classification and categorization of use cases of AI is important in creating laws, since there are many applications and types of AI tools. “For example, the Colorado Division of Insurance has a rule that would regulate AI just for life insurance in Colorado,” Gamino-Cheong said. New York City also has one of the first laws requiring an external AI bias audit, creating a regulatory complex environment for businesses.

Boundaries & Ethics in AI

The panelists discussed the implications of generative AI in government agencies, data privacy, and copyright protection. They acknowledged the potential benefits of using AI to improve efficiency in reviewing applications, credit scores, and other processes, but also recognized the need for a human in the loop to ensure accuracy and trust.

They discussed the challenges in setting boundaries around the use of AI in political advertising, education, and personal information, as well as the potential for AI to be used inappropriately in HR and hiring processes, recidivism prediction, and foster care settings. Applying AI in these cases should not be used for the “end-to-end problem of actually making the prediction of the recommendation directly on whether a candidate should be advanced,” said Vlad Eidelman, chief technology officer at FiscalNote. “I think there’s a lot more nuance there, and there’s potentially a lot more snake oil that can seep in where people want a simple solution.”

The panelists also touched on the need for clear legislation and ethical standards in the use of AI, especially regarding data privacy, copyright protection, and in social media. Beyer said that, in general, there is a need to “train the American people and people around the world to be skeptical of what they see, to look for additional sources, to go deeper, and on the challenge of interpretability on AI, which, for me, is one of the most difficult things to figure out how to regulate.”

How FiscalNote’s AI-Driven Intelligence Can Help Your Team

As AI develops, it’s essential for government affairs teams across industries to monitor policy changes and the conversations surrounding this technology. FiscalNote’s patented AI technology empowers you with the information and tools you need to make smarter decisions and navigate complex situations. Our Application Programming Interface (API) capabilities allow our database of global legislative and regulatory data, elected officials, and government contacts to be seamlessly integrated into your internal processes in a secure manner.