Pony.ai's Driverless Leap, Japan's AI Bill, and Global AI Security Moves - AI Policy Brief #11

Hi there,

Welcome to this week's edition of the AI Policy Brief, where we bring you the latest updates on AI regulations, safety standards, and government policies from around the world. This week, we're looking at new legislative efforts in the United States aimed at addressing AI discrimination, as well as a significant development in China where Pony.ai has received a permit to operate driverless robotaxis in Shenzhen. Meanwhile, Japan is proposing new amendments to its AI bill and data laws, signaling a shift towards more comprehensive AI governance.

In the realm of AI safety, the OWASP GenAI Security Project has been elevated, highlighting the growing focus on securing AI technologies. Additionally, the Spanish police have halted the use of the Veripol AI tool, and a report from the ADL has found bias in AI models. These stories underscore the ongoing challenges and advancements in ensuring AI systems are safe and equitable for all users.

Pony.ai has received a permit to operate fully-driverless robotaxis in Shenzhen's Nanshan District. The Japanese Government plans to introduce an AI Bill and amend data laws to promote AI development.

Georgia lawmakers are advocating for AI regulations to prohibit AI-generated child pornography and election deepfakes. The SEC convened a roundtable on March 27, 2025, to address AI advancements and risks in the financial industry.

Kela reports a 219% increase in dark web discussions about malicious AI tools in 2024. Anduril Industries expands manufacturing in New South Wales to support the Australian Defence Force. Immersive Wisdom, Inc. uses its Digital War Room software in US military exercises. Cyberhill Partners collaborates to enhance DoD biosurveillance with AI.

Denver has issued an RFP to find AI solutions for improving city services. Elliptic introduced new data solutions for the US Government, including enhanced APIs and AI-generated data sets. HCLTech launched a Public Sector Solutions subsidiary to deliver AI-driven technology solutions to US government agencies.

OWASP has elevated its GenAI Security Project to address AI security risks globally. The Spanish National Police have ceased using the Veripol AI tool due to court validity concerns. The ADL report finds bias in AI models, urging developers to address these issues. Zelim installs its AI detection system on the Ambition cruise ship. CDI and Purdue University partner on AI security research. Common Sense Media appoints Bruce Reed to lead AI safety initiatives.

Anthropic has won an initial ruling in a copyright case against music publishers, while a federal court ruled that AI-generated art by Stephen Thaler's 'DABUS' cannot be copyrighted. Anthropic is also seeking dismissal of a separate copyright lawsuit, and OpenAI's new AI image tool is raising copyright concerns.