California amended its AI accountability bill, SB-1074, in response to tech industry criticism, aiming to protect whistleblowers and allow state intervention in potential AI-related crises.
California lawmakers have modified a contentious law that would require AI companies to be responsible for the harm their products cause in response to industry criticism.
Nevertheless, it is uncertain whether the state’s concessions will be sufficient to persuade those who remain skeptical in the technology sector.
SB-1074, or the “Safe and Secure Innovation for Frontier Artificial Intelligence Models Act,” is a bill that aims to safeguard whistleblowers and to allow the state of California to intervene in the event of an AI-related catastrophe.
Cointelegraph recently reported that the tech industry is in a state of disarray due to the bill’s stipulations, with numerous tech luminaries cautioning that it would stifle innovation and growth.
Stifling Innovation
The bill’s original form would have permitted the state to pursue legal action against companies for negligence regarding inadequate safety protocols, regardless of whether the violations led to a catastrophic incident.
Additionally, it would have established a government oversight commission that was accountable for the implementation and enforcement of the aforementioned safety practices.
After receiving negative feedback from the tech industry, including an exhaustive list of suggestions from AI firm Anthropic, California Senator Scott Weiner asserted that his office was under the impression that they had discovered a suitable compromise:
“We accepted a number of very reasonable amendments proposed, and I believe we’ve addressed the core concerns expressed by Anthropic and many others in the industry,”
Central Issues
Nevertheless, it is evident that lawmakers have yet to resolve the primary concern of the tech industry, as evidenced by the bill’s revised language.
Whether the state has the authority to file a lawsuit prior to a catastrophe or solely issue injunctions is tangential to the fundamental issue of liability.
Under SB-1074, the state would still have the right to hold any AI developer accountable for any harm that their products caused.
It is nearly impossible to anticipate all the potential ways in which “harm” could be manifested by an AI model in numerous instances.
The bill’s subsequent phase is a final assembly vote, which, as per the schedule, must occur before August 31. Technology firms operating in California will encounter a regulatory environment that is unparalleled in the nation, unless the governor elects to veto it.