US President Donald Trump on Tuesday, September 29, held a meeting with the heads of leading technology companies working in the field of artificial intelligence. He called the event "an extraordinary gathering" of minds who are "building America's Gilded Age."
Among the guests were the heads of OpenAI, Anthropic, Nvidia, SpaceX, Meta and Google - key figures in the American artificial intelligence industry.
The meeting came amid calls from a number of experts and representatives of the technology sector for greater control over AI agents due to concerns about the potential risks of their further development.
On the day of the meeting, OpenAI, the creator of ChatGPT, confirmed that it would not release its new GPT-6.1 Astra artificial intelligence model due to security concerns. This decision followed a statement by Australian authorities about cases of hacking of government websites and systems by AI agents created by OpenAI.
However, after the meeting, Donald Trump once again tried to assure the public that the threats from the use of new AI models are not so great and companies working in the field of artificial intelligence can take care of the safety of their products.
Here are three main takeaways from Trump's meeting with tech companies at the White House.
1. Agreement on “self-regulation” in the field of AI
After the meeting, Trump, along with technology company executives, answered questions from reporters for about 30 minutes. The president said the meeting participants signed a document designed to serve as a “protection tool” against potential risks associated with AI. At the same time, according to him, the obligations specified in this document are rather of a “moral” nature.
Under the terms of the agreement, the text of which was later published, companies are responsible for ensuring the security of the technologies they create. The companies also agreed to implement protective mechanisms to ensure that AI models operate in accordance with specified principles. These mechanisms must also promptly identify and correct any problems.
They will also work with “independent auditors” to assess whether their AI systems are functioning properly. Independent reviewers are also expected to be required to ensure that AI agents and platforms do not hack “technical systems or gain access to them through unauthorized means.”
The agreement was signed by Donald Trump, Google CEO Sundar Pichai, Anthropic CEO Dario Amodei, Meta Platforms CEO Mark Zuckerberg, OpenAI President Greg Brockman, Elon Musk (SpaceX) and Nvidia CEO Jensen Huang.
“In some ways, it's like the Constitution,” Trump said of the agreement. “It was signed by the most powerful people in the world, and I signed it as president.”
Trump added that he would create a board to oversee the safety of artificial intelligence tools. He did not specify who would be part of it.
Zuckerberg said the agreement gives "the American people and customers confidence that the technology works" as intended and called the meeting a "historic conversation."
Amodei called it a way to "win safely," citing the US president's recent remark that "whoever wins the AI race will win everything."
However, a number of experts noted that this agreement is unlikely to protect people from threats that will arise as AI technologies improve.
José Miguelito Henriques, an artificial intelligence management researcher at Nanyang Technological University in Singapore, said the paper shows an "emerging consensus" but does not address the question of whether AI developers should be held accountable for incidents that occur.
University of Sydney law professor Kimberley Weatherall said the agreement was "completely underwhelming" and came despite calls from some world leaders and industry representatives for greater regulation of new AI technologies.
Weatherall noted that the document appears to allow companies to determine for themselves what is considered safe, and does not stipulate consequences for possible omissions or violations regarding the safety of technologies.
“This agreement should be ignored as an empty distraction,” she concluded.
Trump has repeatedly said that the risks associated with artificial intelligence are not that great. He argued this even as leading companies in the field called for greater oversight of AI technology.
On Monday, OpenAI announced it was discontinuing its latest flagship AI model, GPT-6.1 Astra, due to security concerns. The company's security practices have come under scrutiny following a number of high-profile incidents involving its technology, including instances where AI models it created gained unauthorized access to government sites and systems in Australia.
Ginny Paterson, from the Center for Artificial Intelligence and Digital Ethics at the University of Melbourne, noted that a focus on self-regulation could lead to companies implementing "only minimal protective measures." According to her, the agreement inspires “extremely little confidence”, since recent cases of “AI agents going out of control” indicate an undisciplined approach by companies to the issue of security.
2. Focus on expanding the data center network
Trump reiterated his support for the rapid expansion of data centers, the critical infrastructure that powers artificial intelligence systems.
Plans to roll out data centers in the US have sparked protests from local communities across the country over concerns about the high energy consumption of such facilities and their potential environmental impact. In addition, the construction of data centers could become a political problem for Trump's Republican Party ahead of the November US midterm elections.
According to a poll conducted by Marist Poll in collaboration with NPR and PBS News earlier this month, 65% of registered voters opposed building data centers in their regions.
In response to these concerns, Trump said tech companies will try to keep people "happy" with building data centers.
“These large, powerful, very rich, very smart companies will make enormous contributions to local communities,” Trump told reporters.
Trump previously wrote on social media that the only reason local communities could oppose data centers is if they “want to remain backward and poor.”
Jensen Huang of Nvidia, which produces the AI chips needed to power artificial intelligence systems, said after a meeting with Trump that data centers should be called “superintelligence factories.” These facilities contribute to the "reindustrialization of the United States" and create jobs in various sectors of the economy, he added.
3. Now this is “superintelligence”
Trump also signed an executive order on Tuesday directing US government departments and agencies to begin using the terms "SI" (SI is an acronym for Super Intelligence) and "Super Intelligence" and to stop using the term "artificial intelligence."
According to Ginny Paterson, since representatives of the technology sector have agreed on this term and use it, the United States may well set a new standard in the name of technology.
However, after Trump’s speech at the UN, some experts emphasized that the term “superintelligence” is already used in science, but has a completely different meaning - an intelligence that is superior to humans in everything. They noted that using the term as intended by Trump could create confusion.
