BSA Urges Action Focusing on High-Risk AI

WASHINGTON – The pace of legislation introduced in US state legislatures regulating artificial intelligence (AI) continues to increase, according to a new analysis prepared by BSA | The Software Alliance.

Through February 7, state lawmakers nationwide had introduced a total of 407 AI-related bills across 41 US states where legislatures were in session. Nearly half of those bills (211 bills) were introduced in January 2024 alone, reflecting the sustained interest in AI issues among state lawmakers nationwide.

“There is understandable and significant interest in artificial intelligence policy among state lawmakers, as demonstrated by the big increase in legislation introduced related to AI,” said Craig Albright, Senior Vice President for US Government Relations at BSA. “There are many issues under consideration, but the enterprise software industry is urging state legislatures to coalesce around an approach to addressing high-risk uses of AI that would require companies to have risk management programs and use impact assessments to mitigate risks of bias or discrimination.”

“Some of the legislators who led the efforts to pass comprehensive privacy legislation are now interested in leading on high-risk AI,” Albright continued. “There is an opportunity for a leading state to set the model for AI policy. The challenge for state officials is getting the policy right, combining efforts to combat AI risks with promoting innovation and broad technological adoption.”

According to BSA’s analysis of state legislative action on AI through Feb. 7, 2024:

• There are a total of 407 AI-related bills currently active across 41 US states with legislatures currently in session.
  • Those 407 bills include legislation on a variety of AI-related issues, ranging from mitigating risks of bias and discrimination, establishing task forces, to addressing concerns about deepfakes.
  • 192 of the 407 active bills have to do with deepfake issues.
  • The 407 number includes legislation that was introduced in 2023, and which has carried over into 2024.
  • By comparison, 67 AI-related bills had been introduced by Feb 7, 2023, a 6.1x explosion in a year's time.
• In January 2024:
  • State lawmakers introduced a total of 211 AI-related bills in January 2024 alone.
  • 101 new deepfake bills were introduced in January – meaning that more than half of the 192 deepfake bills under consideration across the country (through Feb. 7) were introduced during the first month of 2024.
  • That averages out to about 50 AI bills per week, including 25 deepfake bills per week in January 2024.
• Of the 41 states in session as of February 7, 40 states had at least one AI-related bill introduced.
• States with most AI-related bills: New York (65), California (29), Tennessee (28), Illinois (27), New Jersey (25).
• Four states had some sort of executive action relating to AI in January (Maryland, Massachusetts, Virginia, Washington).

BSA’s analysis was prepared using the same methodology as an update on AI legislation published in September 2023.

Said Albright: “A strong national law on AI remains the best way to provide regulatory clarity for companies, customers, and governments, and BSA is working with Congress to encourage action at the federal level. Workable laws on AI help promote trust in technology and spur the adoption of AI systems that widely spread the benefits of AI across businesses and communities.”