Artificial intelligence and machine learning are barnstorming their way through industries, and a Georgia study committee is looking into what actions lawmakers might take when they return to session in January.
“How do we look forward, and without stifling innovation, put those proper guard rails in place, if you will, to do that?” state Sen. John Albers, R-Roswell, asked during the fourth meeting of the Senate Study Committee on Artificial Intelligence.
Albers said lawmakers agree that we would love to have one uniform standard for all 50 states but noted that’s not likely to happen anytime soon.
“We do believe that rushing to regulations can have real impacts on the innovative work, and we highly encourage achieving a delicate balance between safeguarding against potential harms and encouraging innovation,” Renzo Soto, TechNet’s executive director for Texas and the southeast, told committee members.
In Texas, the Department of Insurance has notified regulated insurers that they are responsible for the accuracy of the data they use for claims, ratings and underwriting. The agency has explicitly stated that data in AI systems and tools used by insurers are subject to that duty of care.
In transportation, for example, Delta Air Lines created a customer service tool using Amazon bedrock that’s able to navigate all sorts of questions from folks from simple to complex.
“So, as I hope you’ve seen, there’s a lot of potential with AI, and to secure this potential, our members are taking the lead in responsible development and deployment,” Soto said. “We do recognize, however, that there are genuine risks, and they need to be addressed, and that’s why there are already strategies in place by industry to proactively address those issues.
“…We do believe that rushing to regulations can have real impacts on the innovative work, and we highly encourage achieving a delicate balance between safeguarding against potential harms and encouraging innovation,” Soto added.
Soto said policymakers should work to make data privacy and cybersecurity proposals interoperable and also highlighted the importance of leveraging existing authorities already afforded by state regulations and rules that are on the books.
Soto urged lawmakers to:
- Avoid blanket prohibitions on artificial intelligence, machine learning, or other forms of automated decision-making.
- Reserve any restrictions only for specifically identified use cases that present a clearly demonstrated risk of unacceptable harm and narrowly tailor those requirements to the arms identified.
- Not to force developers or deployers of AI or ML to share information that is either proprietary or protected.
- Not to require an AI registry.
- Ensure the safety and security of information by ensuring data retention requirements are appropriately scoped to need and clearly defined by law.
