California is taking a bold step to regulate artificial intelligence used by its government. The state has introduced new rules that require AI vendors to prove they have safeguards against bias, illegal content, and civil rights violations. Companies that fail to meet these standards could be barred from doing business with the state. The move aims to ensure AI technology used in public services is safe, fair, and respects people’s rights.

New Certification Requirements for AI Vendors

Governor Gavin Newsom signed an executive order directing the Department of General Services and the California Department of Technology to create new vendor certifications within 120 days. These certifications will require companies to attest that their AI systems have measures in place to prevent the exploitation or distribution of illegal content, such as child sexual abuse material and non-consensual imagery. Vendors will also need to demonstrate that their models do not display harmful biases and are governed to reduce such risks.

The order emphasizes the importance of protecting civil rights and liberties. Vendors will have to confirm their AI systems do not violate rights related to free speech, voting, human autonomy, or protections against unlawful discrimination, detention, and surveillance. If companies cannot prove they have these safeguards, they risk losing access to lucrative state contracts. This new vetting process aims to hold AI providers accountable for ethical standards in government projects.

Reforming Vendor Eligibility Beyond Finances

Beyond certification, the order directs California’s Government Operations Agency to recommend reforms that could exclude entities found to have unlawfully undermined privacy or civil liberties. This is a broader approach than usual criteria like financial stability or past performance. The goal is to prevent companies linked to civil rights violations from participating in state contracts altogether.

Interestingly, the order grants the state’s Chief Information Security Officer (CISO) the authority to review federal supply chain risk designations. If the CISO believes a federal ban on a company is unwarranted, they can work with other agencies to create guidance that allows continued procurement. This override mechanism reduces geopolitical risks for vendors, protecting those who might otherwise be rejected due to federal sanctions. It also signals California’s intent to set higher ethical standards for AI, even if federal oversight is weaker or changing.

California’s move comes at a time when federal AI regulations are retreating. President Trump recently revoked President Biden’s 2023 executive order aimed at ensuring safe and trustworthy AI development. This rollback eliminated mandatory testing and oversight for high-risk AI models. California’s new rules could fill the gap by establishing state-level standards that prioritize safety and ethics in AI use for government services.

Overall, California’s new policies aim to position the state as a leader in responsible AI development and deployment. With many top AI companies based in California, these rules could influence global standards. The state’s approach signals a serious commitment to protecting civil rights and ensuring that AI benefits society without causing harm or bias.

Inspired by

• https://www.computerworld.com/article/4152400/california-to-bar-ai-vendors-that-cant-prove-bias-safeguards.html

Sources

• shutterstock.com
• ca.gov
• ca.gov