AI oversight was the focus of CBS LA’s reporting featuring Dr. Chris Mattmann, Founder and President of Mattmann.AI, following Governor Gavin Newsom’s announcement of a new executive order addressing artificial intelligence safety.
In the September 18 report, CBS LA political reporter Tom Wait examined California’s approach to independent evaluations, AI safety, and potential emergency shutdown capabilities. Mattmann discussed why outside scrutiny matters and the challenges of understanding how AI models are developed.
What California’s AI Executive Order Addresses
The executive order accelerates implementation of California’s independent AI oversight laws and convenes experts to recommend additional safeguards.
Measures under consideration include independent monitors within frontier AI laboratories, verification of safety frameworks and risk assessments, and emergency shutdown capabilities commonly described as an AI “kill switch.”
The order advances the evaluation of these proposals. It does not immediately require every AI company to install a shutdown mechanism.
Read the Governor’s announcement about the executive order.
Independent AI Oversight and Human Alignment
Mattmann highlighted the importance of allowing external inspectors to examine AI development, including the process known as alignment.
He described alignment as the stage at which developers teach AI systems human values. Independent examination of that work can help bring greater scrutiny to the expectations and safeguards built into increasingly capable models.
The discussion connects a broad policy debate to a practical question: how can people outside an AI company assess whether its development processes and safety claims hold up?
Training Data Transparency and Model Provenance
Mattmann also addressed the difficulty of learning what data companies use to train their models. He criticized the reliance on litigation to uncover information about model provenance—the origins and development history of an AI system.
In his comments, Mattmann argued that relying on lawsuits to obtain that information is an inadequate foundation for transparency and public understanding.
These concerns place training data, development practices, and access to evidence at the center of the AI oversight conversation.
Watch the CBS LA Report
Watch CBS LA’s coverage featuring Dr. Chris Mattmann below. His comments begin at approximately 1:36, with further discussion of training data transparency at approximately 2:46.
Watch the full CBS LA report on YouTube.
Thank you to Tom Wait, Pat Harvey, and the CBS LA team for covering these important questions.
For AI strategy, executive advisory, speaking engagements, and media inquiries, visit Mattmann.AI.