This Week in Government Technology – July 21st-28th, 2024

This week’s highlights explore critical advancements and regulatory efforts in the AI landscape. California’s proposed “Safe and Secure Innovation for Frontier Artificial Intelligence Models Act,” led by state Senator Scott Wiener, aims to regulate AI development through the Frontier Model Division and CalCompute program. While supported by AI safety groups, it faces opposition from tech giants like Google and Meta, who argue it could stifle innovation. In response to this bill, tech leaders express concerns about its potential impact on startups and liability for AI misuse. Meanwhile, Washington state has established an 18-member task force to balance AI innovation with proper oversight, discussing topics like intellectual property and bias. Additionally, Apple joins other tech giants in adopting the Biden administration’s voluntary AI safeguards, testing AI systems for security and discrimination risks. These efforts underscore the ongoing challenge of balancing AI’s transformative potential with its significant risks.

California’s Legislative Push for AI Safety and Security

California’s proposed “Safe and Secure Innovation for Frontier Artificial Intelligence Models Act,” spearheaded by state Senator Scott Wiener, aims to regulate the development and deployment of advanced AI models. The bill seeks to preemptively address the risks associated with AI by establishing the Frontier Model Division, a regulatory body, and the CalCompute program to support large-scale AI model development. While the bill has garnered support from AI safety advocacy groups, it faces opposition from tech giants like Google and Meta, who argue it could stifle innovation and disproportionately burden AI model developers. The bill underscores the urgent need to balance AI’s potential benefits with its significant risks, including misuse of autonomous weapons and cybersecurity threats.

Tech Industry Responds to California’s Proposed AI Legislation

In response to Senator Scott Wiener’s “Safe and Secure Innovation for Frontier Artificial Intelligence Models Act,” tech leaders express concerns about potential negative impacts on innovation. The bill aims to regulate AI development to prevent misuse but has been criticized by figures such as Garry Tan of Y Combinator and Andrew Ng of Google Brain for potentially stifling AI advancements. Despite amendments to address industry feedback, critics argue the legislation could hinder startups and impose unfair liabilities on developers for third-party misuse of their AI models. Wiener remains open to further revisions, emphasizing the bill’s narrower scope compared to the European Union’s AI regulations.

Washington’s New Task Force Takes Action

Washington state has established an 18-member task force to address the growing influence of artificial intelligence, with its first meeting scheduled for Friday. The task force, consisting of state lawmakers, tech industry leaders, advocacy group representatives, and government officials, aims to balance fostering AI innovation with ensuring proper regulation. Topics for discussion include intellectual property, AI oversight, bias, and racial disparities, transparency, and support for AI advancements. Attorney General Bob Ferguson highlighted Washington’s leadership role in technology and the importance of developing policies that prioritize public interests. The task force will meet biannually over the next two years, culminating in a policy recommendation report by July 2026.

Apple Joins Tech Giants in Adopting Biden’s AI Safeguards

Apple Inc. has agreed to adopt the Biden administration’s voluntary AI safeguards, joining other tech giants like Amazon, Microsoft, and OpenAI. These measures involve testing AI systems for security flaws, discriminatory tendencies, and national security risks, with the results to be shared transparently with governments, civil society, and academia. This move comes as Apple plans to integrate OpenAI’s ChatGPT into its iPhone voice-command assistant. While the guidelines are not enforceable, they represent an effort by the administration to encourage responsible AI development amid growing concerns about AI’s potential dangers and discriminatory impacts.