FOR IMMEDIATE RELEASE
September 23, 2026
EqualAI President & CEO Miriam Vogel testified at an Emergency Hearing Calling for U.S.-China Agreement organized by Congressman Ro Khanna, Ranking Member of the United States House Select Committee on Strategic Competition between the United States and the Chinese Communist Party, as covered by Fox Business.
Below is the text of her full written testimony.
Miriam Vogel Testimony
The meeting between President Trump and President Xi creates a timely opening for progress on AI governance, and we are grateful for the chance to be part of that conservation.
I have helped companies and policymakers translate principles into practical guardrails for the development and deployment of AI for the last eight years. Last October, I co-authored Governing the Machine, an award-winning book on AI governance. I have seen what happens when institutions get ahead of emerging risks—and when they do not.
From that perspective, I would offer three points.
First, effective AI governance requires international engagement —including with China.
There is broad agreement that AI development has outpaced the governance frameworks needed to manage it. If the US were to pause or restrict frontier AI development without China at the table, we could constrain American companies while leaving a major source of risk unaddressed.
President Reagan’s approach to arms negotiations with the Soviet Union “trust, but verify” recognized the need for both engagement and verification. AI governance will require the same combination: mechanisms that make commitments transparent and verifiable.
There are areas where our interests overlap: preventing catastrophe, protecting children, and reducing the possibility that increasingly capable systems create harm beyond our intent or imagination.
The American public wants the government to engage. In an EqualAI commissioned YouGov survey of 2,000 Americans, a majority wanted Congress to safeguard AI, particularly, our infrastructure, data, and children. Only 4 percent wanted Congress to abstain.
This moment presents an opportunity to begin building mechanisms for dialogue around risks neither country can manage alone.
Second, American leadership on AI requires leadership in AI governance.
Leadership is not only about building the most capable models. It is also about establishing the frameworks that allow those systems to be trusted and adopted.
China has taken important steps toward governance with regulation of algorithmic recommendation systems and generative AI, and in April 2026, launched an “AI + Education” action plan designed to build AI literacy across all school levels by 2030.
Our institutions and values are different but if the US wants to shape global AI norms, we first have to define and operationalize our own.
Effective governance does not slow innovation. It is the infrastructure that allows innovation to scale.
We fly 45,000 flights across U.S. airspace daily because passengers trust international safeguards for certification, inspection, maintenance, and investigation. We put our families in vehicles because we know they’ve met established global and national safety standards.
AI needs that same institutional discipline.
And we need an AI-literate workforce. This is a workforce issue. A competitiveness issue. And a national security issue.
Third, we need governance throughout the AI lifecycle.
Too often, proposed safeguards end with the model development.
Some of the highest-stakes AI interactions occur during deployment: in financial institutions, hospitals, workplaces, and public institutions, where governance can be weakest. The World Economic Forum reported that less than 1% of companies have strong AI governance, McKinsey reported that less than a third have any governance in place.
Consider agentic AI: systems that can take actions, interact with other systems, and make decisions with increasing autonomy.
At EqualAI’s recent Agentic AI Governance Summit, we simulated what happens when agentic systems operate without sufficient governance. Ordinary deployment quickly escalated into incidents—and then crises.
It reinforced EqualAI’s core governance practices: Organizations need visibility on where AI is being used, what it can access, who owns it, what decisions it can make, and what happens when it fails.
We also need independent evaluations before deployment. Our survey found 63 percent of Americans, across party lines, trust independent scientists most as external validators of AI safety. The institutional model is open to debate. But the fundamental questions are clear: who independently evaluates frontier systems, against what benchmarks, and with what consequences?
And governance cannot end at deployment. We need certainty and alignment on expectations for each of the points I noted, including a system for AI incident reporting: clear definitions, reporting pathways, and appropriate protections for companies that report significant incidents. An NTSB-like model could create infrastructure to share lessons and help prevent repeat failures.
To close: Discussions about a pause in AI development must include China. American leadership requires governance that allows AI systems to earn and deserve trust. And safeguards should extend across the AI lifecycle.
We have navigated technological transformation before, not by stopping innovation, but by building the institutions capable of governing it. That is what this moment asks of us again.
Thank you,
Miriam Vogel
President and CEO
EqualAI
