The outgoing presidential artificial intelligence adviser said that the Trump administration will not introduce a formal artificial intelligence licensing regulatory system; although the White House is using emergency executive powers to suspend the release of many top large models. In his first in-depth interview since resigning from the government last month, Sriram Krishnan said: "The United States will not establish a dedicated regulatory agency for artificial intelligence similar to the U.S. Food and Drug Administration (FDA)."
He added: "This administration and the president have firmly opposed cumbersome, harsh and bureaucratic red tape since the beginning of taking office. We will not act as industry referees and artificially select industry winners and losers."
Krishnan said that if a centralized regulatory agency is established and companies are required to form legal teams before they can release AI models to the outside world, it will only hinder the innovation process of the artificial intelligence industry. "This type of regulatory model will never take off during President Trump's term."
Just a few weeks before this statement, the U.S. government launched unprecedented intervention measures on the grounds of national security, forcing Anthropic to withdraw its most powerful large model Mythos and suspending the launch of the OpenAI 5.6 large model.
Krishnan, who was born in India, worked as a venture capitalist and worked with Elon Musk before entering government. Together with David Sachs, the White House AI chief, he is an advocate within the Trump camp.Loose and prudent AI regulationcore characters.
Some officials in Washington believe that the current administration's laissez-faire regulatory orientation has triggered resistance to artificial intelligence among American voters and Trump's "Make America Great Again" (MAGA) supporters. People are generally worried about the disorderly expansion of data centers and the various negative social impacts of AI.
Data from Data Center Watch, an industry research organization, shows that the vast majority of Americans support the introduction of strict regulations to restrict the development of artificial intelligence; in the first three months of 2026, at least 75 data center projects in the United States with a total investment of approximately US$130 billion were shelved due to opposition from local people.
Krishnan has pushed for the rejection of AI regulatory rules issued by various states, and took the lead in drafting the White House framework document, which only formulated narrow-scope laws around the protection of minors and content compliance. In his view, the public’s resistance to AI is rooted in the doomsday pessimism spread by the artificial intelligence industry itself.
Krishnan said that the AI industry has failed to fully explain to the public the benefits brought by the technology, such as the application value of AI in the field of precision medical diagnosis.
He said that the heads of major AI laboratories in the United States have over-exaggerated the pessimistic future of technology: various dystopian narratives, ranging from mass unemployment to human survival-level risks, have emerged one after another, which has also made many people resist: "I'm really not sure whether I need this technology."
In his view, voters hope to participate in the formulation of AI rules and do not want to see a few technology giants use AI to become powerful and seize huge wealth. This public opinion has been accurately grasped by Trump. Recently, Trump proposed that technology companies sell part of their equity to all Americans, and has communicated with OpenAI CEO Sam Altman about this plan.
Although many people in Silicon Valley warned that the plan was essentially a disguised nationalization, Krishnan still expressed support for it.
"The form of equity implementation can be discussed in many ways, but ordinary people should enjoy the dividends brought by the development of AI. When the public uses large models and sees the revenue of technology companies rising, they can personally feel the benefits from it. This is a reasonable development model."
He believes that the public must be made aware that artificial intelligence is a tool that empowers ordinary people.
Krishnan, 42, joined the White House a few days after Trump took office for his second term in 2025. He is also one of the few industry figures in Silicon Valley who settled in Washington early to help the government build an AI governance framework.
Talking about his special experience in Washington as a bridge between Silicon Valley and the political world, he revealed that he once explained to a group of cabinet officials "Rocco's Basilisk", a classic thought experiment that explores the revenge of super artificial intelligence in the future.
In the past 18 months, Krishnan and Sachs have joined forces to push states to abolish strict AI safety regulatory rules and only adopt lightweight monitoring of national security and network security risks of cutting-edge large models.
However, the recent emergence of a number of new large-scale models with super network attack and defense capabilities has once again made the US government's call for strengthening AI control stronger.
Altman publicly stated last week that the U.S. government's request to postpone the launch of OpenAI's latest 5.6 version model is not the best option; Silicon Valley investment institutions also warned against temporarily establishing a licensing regulatory mechanism that requires cutting-edge large models to be approved by the White House.
Some believe that the Trump administration’s move is intended to retaliate against Anthropic. Trump once accused the company of belonging to the “radical left camp” because it refused to allow the U.S. Department of Defense to use AI technology in the military field. Krishnan denied this.
In the past few weeks, Krishnan has successively explained to leaders of many countries the United States’ decision to unilaterally suspend the Mythos model’s external services. He supports Trump's use of export controls to force Anthropic to temporarily remove the model, but said the government made this decision out of desperation: Amazon had previously discovered high-risk security vulnerabilities in the model and had to intervene out of national security considerations.
Krishnan said that the U.S. government will definitely make every effort to protect its systems and national security from threats. Vice President Vance and other top government officials have repeatedly called for protecting critical infrastructure from potential threats from cutting-edge AI models.
From the perspective of long-term governance, Krishnan advocates that the supervision rights and responsibilities of cutting-edge large models should be borne by the industry independently. A voluntary third-party clearing agency for the industry can be established to attract the participation of large technology companies, chip manufacturers, and network security companies, and collaborate with the intelligence department and the Ministry of National Defense to jointly investigate model security vulnerabilities.
This regulatory model will be in line with the executive order previously signed by the president: before companies release cutting-edge large models, they must voluntarily submit them to the U.S. government for a 30-day security review.
However, Krishnan also warned that if the government puts aside the implementation of cutting-edge AI technology for a long time, it will severely damage the competitiveness of U.S. technological innovation.
Months after Sachs stepped down as White House AI chief, Krishnan also chose to leave the government. He is currently recruiting AI technical engineers to assist European and American governments in formulating artificial intelligence regulatory policies. He mentioned that there is a high probability that US allies will choose to useopen weight model, this type of model exposes training parameters, and countries can independently fine-tune and optimize based on local scenarios.
When asked whether the future Democratic Party government will use the current government's export control measures to restrict the implementation of AI technology, Krishnan responded: "I will not predict the policies of the next government. I will only focus on the current governance choices of this government."