In summary
In an executive order, California’s governor has pushed back against actions taken by the Trump administration regarding a California AI startup. Concurrently, the governor is implementing more regulations to oversee the technology.
The next time the federal government designates a business as a supply-chain risk, as seen recently with San Francisco AI developer Anthropic, California will independently evaluate that classification before deciding whether to engage in business with the company.
This decision follows an executive order signed by Governor Gavin Newsom on Monday. This order emerged from a dispute between Anthropic and the Department of Defense concerning clauses in contracts that prohibit the military from using Anthropic’s technologies for domestic surveillance and fully autonomous weaponry.
By labeling Anthropic as a supply chain risk, the Department of Defense effectively prevented the startup from competing for specific military contracts and subcontracts. Recently, a judge issued a temporary injunction to suspend this designation.
The overarching goal of Newsom’s order is to create safeguards surrounding the use of AI by state employees while simultaneously promoting the technology’s accelerated adoption.
California hosts many of the world’s leading AI firms and is at the forefront of national AI regulations.
Under the new order, state agencies are required to:
- Formulate recommendations for contract standards related to AI, including its potential to generate child sexual abuse material, infringe upon civil liberties and rights, or violate legal protections against unlawful discrimination, detention, and surveillance. They should facilitate access for employees to “vetted GenAI tools.”
- Revise the State Digital Strategy to identify how generative AI can enhance government transparency and accountability, boost performance, and make services more accessible to all Californians.
- Create generative AI applications that allow Californians to access government services.
- Provide guidelines for state employees on adding watermarks to AI-generated images and videos.
These directives come at a pivotal moment, as over 20 California departments are working on developing and utilizing Poppy, a generative AI assistant for state employees. Meanwhile, several agencies are testing AI for tasks such as assisting employees and providing support to homeless individuals and businesses. This comes as state courts and municipalities are expanding their use of the technology.
Newsom’s office emphasized that the previous administration under President Donald Trump had weakened protections and neglected the potential harms of AI technology.
“Unlike the Trump administration, California is dedicated to ensuring that AI solutions adopted and implemented by [California]… cannot be exploited by malicious entities,” stated the governor’s office in a press release regarding the new order.
On a federal level, Trump has issued executive orders aimed at deterring state-level AI regulations and has urged federal agencies to implement AI to streamline processes and reduce federal regulations, as well as accelerate decision-making in Medicare. The White House recently introduced an AI policy framework that the president intends for Congress to consider. This proposal adopts a lenient regulatory stance and does not tackle issues of bias, discrimination, or civil rights.
This marks the second executive order issued by Newsom to address AI concerns. A 2023 order focused specifically on generative AI technologies like those powering ChatGPT and Midjourney, similarly calling for increased utilization of AI in state agencies while mandating that safeguards be established.
The approach Newsom is taking towards AI is closely observed by union leaders. Earlier this year, they declared that they will not support his presidential ambitions without stronger worker protections related to AI technology. At the same time, major tech donors are actively investing in influencing California politics ahead of the upcoming midterm elections.
