Weekly Policy Summaries, September 30, 2024:

Veto on California's AI Safety Bill Sparks Debate

Governor Gavin Newsom just axed the AI Accountability and Safety Act (SB 1047), a bill intended to ensure the safety of large-scale AI systems by enforcing stringent regulations and allowing the California Attorney General to sue if these technologies wreak havoc. Newsom pointed out that the bill, which would've been some of the broadest AI regulations in the nation, lacked an evidence-based approach to AI safety, but admitted a "California-only" strategy might yet be necessary.

So what's the ITIF's take on Newsom's veto? Check out their press statement to learn why the California bill would've undercut more sensible federal efforts.

Meanwhile, Newsom put his digital signature on 18 AI bills, among them the AI Training Data Transparency Act (AB 2013) and the California AI Transparency Act (SB 942). The former requires generative AI developers to post info on their site about the data used to train their systems once the tech is available for use in California. The latter makes developers provide detection tools for identifying AI-generated content. Newsom also signed bills that amended existing ones, such as AB 1831, which now includes digitally altered or AI-generated child pornography under existing California statutes covering these offenses.

Over on the federal front, the OMB issued new guidance to help federal agencies smartly acquire AI systems. The memo focuses on three main points: agencies should incorporate AI risk management into their procurement processes, promote collaboration across teams and agencies, and ensure AI purchases promote competition by avoiding vendor lock-in.

Now, let's dive into the enrichment data:

Looming Federal Moratorium: The U.S. House of Representatives has approved a proposal for a 10-year moratorium on state AI laws, as part of a budget reconciliation bill. If passed, this moratorium aims to stop states from enacting their own AI regulations, favoring a federal preemption over state laws.[1][5]
Federal AI Modernization: The House budget reconciliation package includes provisions to modernize federal IT systems using commercial AI. This includes a $500 million allocation to the Department of Commerce to integrate AI into federal systems, enhancing efficiency and cybersecurity.[1]
AI Governance Debate: There's a heated discussion about AI governance, with some pushing for federal oversight to unify AI regulations across the U.S., while others argue that a moratorium could stifle consumer protection and state-level innovation.[2][4]

These developments signal a trend towards federal involvement in AI regulation, but they don't address California's specific legislation or veto directly. Keep an eye on this evolving situation!

The veto of the AI Accountability and Safety Act in California has sparked debate about the appropriate level of regulation for AI technology, particularly in light of ongoing discussions about federal AI governance.
The reception to the California bill's veto also involves consideration of its potential impact on future federal efforts to enforce AI safety regulations.
The International Technology and Innovation Foundation (ITIF) criticized Newsom's veto, arguing it undermines more sensible federal efforts to regulate AI.
The OMB's guidance for federal agencies to smartly acquire AI systems demonstrates increased federal interest in AI regulation, creating a contrast with California's recent veto of its AI Accountability and Safety Act.