Trump rejects AI giants’ warnings of existential threat
US President Donald Trump has brushed aside mounting warnings from leading artificial intelligence executives that increasingly powerful AI could pose an existential threat to humanity, insisting that Washington’s priority should be defeating China in the technological race.
Speaking to reporters in Ireland on Sunday, Trump acknowledged that some safeguards could be introduced but suggested the dangers were being exaggerated by unspecified “negative forces.”
“We’re leading China in AI, we’re the most sophisticated country in the world, and frankly I want to keep it that way… Whoever wins AI wins,” he added, arguing that critics were raising scenarios that “won’t happen.”
’Just slow down yourselves’
Trump’s remarks came after Anthropic CEO Dario Amodei called for slowing frontier AI development, warning that increasingly autonomous systems could soon improve their own capabilities, evade existing safeguards and unleash large-scale cyberattacks beyond human control.
In a rare show of unity, his rivals – OpenAI CEO Sam Altman and SpaceXAI CEO Elon Musk – both backed the proposal.
White House AI adviser David Sacks responded by accusing OpenAI and Anthropic of being a “duopoly on frontier intelligence” and acting as though “antitrust law has to be suspended so you can form a cartel.”
“If the unreleased models are scary enough that you think you should slow down, I support your decision to be responsible,” Sacks wrote. “But stop pretending you need anyone else’s permission.”
China’s AI surge defies sanctions
Sacks also questioned whether China would ever agree to a coordinated international slowdown – an issue increasingly central to Washington’s AI debate.
Despite US sanctions and restrictions on exports of advanced chips and other technology, Chinese developers have rapidly narrowed the gap with their American rivals, particularly through cheaper open-source and open-weight models.
A US congressional advisory body warned earlier this year that Chinese open-source AI was developing a “self-reinforcing competitive advantage,” with models from companies including Alibaba, Moonshot and MiniMax dominating usage rankings on platforms such as Hugging Face and OpenRouter.
Moonshot AI in July released Kimi K3, the world’s largest 2.8-trillion-parameter open-weight system, while Alibaba recently released Qwen3.8-Max with 2.4 trillion parameters. DeepSeek and other Chinese developers have also released increasingly capable models that can be downloaded, modified and deployed independently rather than accessed exclusively through proprietary Western platforms.
Rival AI orders
Beijing has sought to turn that open-source approach into a broader geopolitical alternative to reliance on Western-controlled AI infrastructure.
Addressing the BRICS summit in New Delhi this weekend, Chinese President Xi Jinping proposed establishing a BRICS AI open-source community, supporting joint development and deployment of large language models, providing specialized training, and building what he called an “open ecosystem for AI.”
Xi also invited BRICS members to participate in the World Artificial Intelligence Cooperation Organization (WAICO), a China-backed body established in Shanghai in July.
Twenty-nine countries, including Russia, Indonesia, Pakistan and Kazakhstan, signed WAICO’s founding agreement. Beijing says the organization is intended to give developing countries greater access to AI technology and a role in setting global rules, and has pledged thousands of AI training opportunities as well as cooperation centers with BRICS, ASEAN, the African Union and other organizations.
The initiative is emerging as a rival to Washington’s Pax Silica framework, which seeks to organize semiconductor, critical-mineral and AI supply chains among what the US describes as “trusted” partners. Washington is reportedly preparing to warn participating states that they may have to choose between the American-led system and Beijing’s WAICO.
