Trump and tech giants sign AI self-regulation principles document to set up a 10-member committee to monitor industry development

The Zhitong Finance App learned that US President Trump said on Tuesday that he had signed a “morally binding” artificial intelligence (AI) document with tech leaders after holding a luncheon with a number of tech industry executives at the White House. At a time when AI security risks are sparking more and more discussions, the Trump administration still favors dealing with related issues through industry self-regulation rather than suspending technology development, and is considering setting up a 10-member committee to oversee the AI industry.

US House Speaker Johnson described the document as a “statement of principles” that the industry voluntarily abides by, and said that the White House will provide guidance for the development of the AI industry. This arrangement is basically in line with the Trump administration's previous AI policy direction. In June of this year, an executive order issued by the White House proposed a voluntary cooperation framework for cutting-edge models while promoting advanced AI innovation, and made it clear that this framework should not be interpreted as establishing a mandatory government licensing, pre-examination, or approval system.

Trump emphasizes “self-regulation” of the industry and proposes to set up a 10-member committee

Accompanied by technology company executives, Trump told reporters that he has seen the AI industry show “very strong self-regulation”. The government also has existing regulatory and enforcement mechanisms such as the Department of Justice and the Federal Bureau of Investigation (FBI), but the industry's own restrictions are also very important.

Trump said the US government will not stop AI development. Meanwhile, he revealed that the government is considering setting up a 10-member committee to oversee the AI industry and plans to appoint a new “AI director” within the next three to four days.

Trump also expressed support for large-scale, multi-billion dollar data center projects. He believes that these projects are positive for the US, and technology companies also need to let the communities where the data centers are located benefit from the investment.

The AI policy previously announced by the White House also emphasizes speeding up technology development and application. The National Security Presidential Memorandum issued in June this year proposed that the United States will accelerate the adoption of AI in the field of national security, while requiring relevant systems to have reliability, robustness, and controllability, and establish corresponding testing, verification, and risk management mechanisms.

AI security concerns are heating up Anthropic calls for a balance between competition and security

However, while Trump emphasizes self-regulation, there are still obvious concerns within the AI industry about the security risks of rapid technological development. Anthropic CEO Dario Amodei (Dario Amodei) once again discussed AI risks when he left the White House and said that the relevant regulatory rules are still being discussed.

Amodei said that all parties need to work together not only to ensure that the US can maintain its advantage in AI competition, but also to ensure that this process progresses safely.

Recently, the debate between Washington and Silicon Valley over the speed of AI development has further heated up as cases of cyber attacks involving or coordinated by AI agents have increased, and some researchers have warned about the risks that may arise from more advanced AI systems.

Amodei and OpenAI CEO Sam Altman (Sam Altman) have both called for slowing down part of the AI development process, while Trump previously criticized some AI security concerns.

The Trump administration's current official policy is an attempt to find a balance between speeding up innovation and security. An executive order issued by the White House in June requires the establishment of an evaluation mechanism for the network capabilities of advanced AI models, and allows developers to voluntarily provide models to the federal government for safety assessments before the cutting-edge models are officially released. At the same time, it is clear that no mandatory model release licensing system will be established as a result.

OpenAI delays GPT-6.1 Astra release, re-focusing on AI model security

AI security issues have recently received further attention. OpenAI delayed the release of the latest GPT-6.1 Astra model this week due to security issues. Just last week, the company also said it was carrying out an “extensive” review of model behavior.

The report pointed out that after the Hugging Face attack, which attracted widespread attention this summer, OpenAI has already revealed a number of incidents of unauthorized model behavior, further increasing concerns from the outside world that autonomous AI systems may be out of control.

These events have also caused the industry to face an increasingly prominent policy challenge, that is, as AI agents gradually gain the ability to call software, access data, and even perform tasks independently, whether simply relying on the enterprise's own security mechanisms is sufficient to control potential risks.

AMD (AMD.US) CEO Su Zifeng expressed positive views on Tuesday's White House meeting. She said that there was both optimism about AI development and a strong sense of responsibility on the scene. She believes that the responsibilities of technology companies include not only showcasing the potential of technology, but also ensuring technology safety.

Tech giants such as Hwang In-hoon, Musk, and Zuckerberg gather at the White House

The White House luncheon on Tuesday brought together many heavyweights in the US tech industry. According to the announced seating arrangement, Trump was seated next to Nvidia (NVDA.US) CEO Hwang In-hoon and Tesla (TSLA.US) and SpaceX CEO Musk, followed by Meta (META.US) CEO Zuckerberg and Google's parent company Alphabet (GOOGL.US) CEO Pichai.

Opposing Trump is US Vice President Vance, who is flanked by Amazon (AMZN.US) founder Bezos and House Speaker Johnson, respectively. Other participants included senior government officials such as Microsoft (MSFT.US) CEO Nadella, Anthropic co-founder Tom Brown, OpenAI President Greg Brockman, and US Treasury Secretary Bezent.

Alex Karp, CEO of Palantir (PLTR.US), said before the event began that the tech industry must take responsibility for the risks it is aware of, and emphasized that technology companies should not have a different set of rules than ordinary Americans.

AI development and regulation have entered a critical stage, the US government launched an AI government chatbot

The tech industry luncheon was also held at the same time as an all-day event hosted by Trump and Vice President Vance. During the event, the US government unveiled a new AI chatbot for government websites, showing that the government itself is also speeding up the application of AI technology.

The White House has introduced a number of AI policies since this year. While the emphasis is on reducing regulatory barriers that may hinder innovation, it also puts forward new security requirements for cutting-edge model cybersecurity, critical infrastructure, and national security applications. An executive order issued in June requires the government to strengthen AI-driven cyber defenses while also giving priority to cracking down on the use of AI to illegally invade or damage computer systems.

Judging from the information released at Tuesday's meeting, the Trump administration has not changed the overall direction of promoting the rapid development of technology due to recent AI security disputes. The future policy framework is more likely to revolve around corporate self-regulation, government safety supervision, and enforcement of existing laws, and how the newly established AI Supervisory Committee and the upcoming “AI Supervisor” will participate remains to be further clarified.

Meanwhile, as cutting-edge AI laboratories such as OpenAI begin to adjust model release plans due to security issues, differences within the AI industry over “speed of development” and “security boundaries” are becoming more prominent.