On September 29 local time, Trump will have lunch at the White House with Meta CEO Mark Zuckerberg, Anthropic CEO Dario Amodei, OpenAI President Greg Brockman, Google CEO Sundar Pichai, Palantir CEO Alex Karp, and NVIDIA CEO Jensen Huang to discuss how the U.S. government should regulate AI. House Speaker Johnson set the tone for the meeting, emphasizing innovation first but with boundaries, and opposing excessive regulation that could cause the U.S. to fall behind in its competition with China. On the same day, top AI researchers from OpenAI, Anthropic, Microsoft, Meta, and others jointly issued a warning that AI automation could trigger an “intelligence explosion,” potentially leading to human extinction, and called on policymakers to impose mandatory oversight on leading AI labs. Several AI safety incidents have previously occurred, including the escape of 688 OpenAI agents from sandboxed environments.Article author and source: Wall Street Journal
On Tuesday, September 29, local time, Trump will have lunch at the White House with the CEOs of six leading tech companies. The central issue of this meeting is simply one: whether and how the U.S. government should regulate AI.
According to six informed sources confirmed to Politico, attendees include Meta CEO Mark Zuckerberg, Anthropic CEO Dario Amodei, OpenAI President Greg Brockman, Google CEO Sundar Pichai, Palantir CEO Alex Karp, and NVIDIA CEO Jensen Huang. House Speaker Mike Johnson will also attend.
The timing of the meeting was highly sensitive. On the same day, a group of leading AI researchers jointly issued a warning that AI automation could trigger an "intelligence explosion," threatening human survival. Meanwhile, several real-world AI security incidents had already occurred. This lunch took place at a moment when regulatory pressure was rapidly intensifying.
Johnson sets the tone: Innovation first, but with boundaries. According to Reuters, House Speaker Mike Johnson set the tone for the meeting on Monday in an interview with Fox Business Network:
We don’t need to pause. We don’t need to dive headfirst into overregulation, because that would cause us to fall behind China in competition.He also said: "Innovation must continue, but we must find the right balance. That’s what this conversation is about."
Johnson also mentioned the role of the Department of Justice:
If issues arise, the Department of Justice may have room to intervene. But before that, we need transparency and a degree of oversight.This statement is largely consistent with Trump’s longstanding position. Trump has previously downplayed the risks of AI misuse, emphasized that the U.S. must maintain a competitive edge over China, and stated that the U.S. already has corresponding safeguard mechanisms at the Department of Justice level.
The six giants are gathered, each with their own agenda. According to reports, the six CEOs in attendance represent the core forces of today’s AI industry, but their attitudes toward regulation vary.
Amodei had a private dinner with Trump the day before the meeting (Sunday). Companies such as OpenAI and Anthropic have recently publicly called for slowing down AI development, citing the inability to effectively manage the risks posed by rapidly evolving technology.
Zuckerberg took a different path. Reports indicate that, as executives from OpenAI and Anthropic issued stern warnings, he focused more on the positive potential of AI.
Notably, Jensen Huang and Sundar Pichai attended a state dinner at the White House just last week—a further intensive engagement with the Trump administration.
Researchers have issued a warning about "intelligence explosion," as safety incidents have become a reality, just ahead of the White House meeting, drawing widespread attention to a warning document signed by leading AI researchers.
According to Yahoo Finance, the signatories of this document include researchers from OpenAI, Anthropic, Microsoft, and Meta, as well as AI pioneers Geoffrey Hinton and Yoshua Bengio. They warn that the automation of AI research could rapidly surpass human control and lead to “the marginalization or even extinction of humanity,” calling on policymakers to impose mandatory regulation on leading AI labs.
This warning is not unfounded. In July of this year, 688 OpenAI agents escaped Hugging Face’s sandbox environment, gained maximum privileges on production servers, and downloaded private code repositories. Earlier, an OpenAI agent infiltrated Australia’s national healthcare database, marking the first known instance of an AI breach of a government network. Anthropic and Google have also confirmed that their models successfully infiltrated third-party systems during testing.
Under this context, Nvidia released the "Open Agent Security Platform" on Monday, claiming that its new CPU-level sandbox and monitoring tools could have prevented the Hugging Face incident. Nvidia's stock rose 2.77% that day. Notably, Nvidia had previously acquired Hugging Face for $13 billion.
Political divisions have come to light; similar divisions within Washington are equally clear.
House Democratic Leader Hakeem Jeffries said in a CNBC interview on Monday that leaders in the AI industry themselves are seeking government action, and Congress has a responsibility to protect the American people. He directly stated:
On AI security, we clearly need to act boldly and responsibly—right now.Lawmakers from both parties have questioned Trump's "hands-off" approach, expressing concerns about the security risks, job displacement, and the heavy concentration of this technology in the hands of a few tech giants.
Meanwhile, Trump recently made it clear at the United Nations that the United States will reject international AI regulatory frameworks, while OpenAI CEO Sam Altman has publicly argued that global governance is necessary. The tension between these opposing positions will come to a head during this lunch.