A three-day state visit between the United States and China ended on Friday without a major agreement on artificial intelligence, The Guardian reported. The biggest announcement from the summit was a two-month extension of a trade truce between the two countries.
What each side said
China’s president, Xi Jinping, said both countries have a “responsibility to develop and manage AI for good”, and that they should “ensure that the development of AI is always under human control”. China has set itself the goal of next-generation smart terminals and AI agents reaching a penetration of more than 90 per cent by 2030, according to the report.
President Donald Trump, for his part, pushed to rebrand AI as “super intelligence”. Several US technology leaders, among them Elon Musk, AMD’s Lisa Su, Nvidia’s Jensen Huang and Apple’s Tim Cook, attended the state dinner; their Chinese counterparts were not present, the paper noted.
What observers expected, and what they got
Some had hoped for more. Representative Ro Khanna, who had proposed technical working groups between the two countries and a pause on self-improving AI, told The Guardian: “It was such a layup.”
The only thing they seemed to agree about is that neither of them wants to slow down with this technology.
Jonathan Czin, Brookings Institution, quoted in The Guardian
The two leaders are set to meet again at the APEC summit in Shenzhen in November and at the G20 in December, so the question will come back. For now, there is no shared rulebook between the two largest AI powers.
Why this matters on your desk, not just in Beijing and Washington
Set the geopolitics aside and one thing is clear from this summit: neither of the world’s two AI superpowers is planning to slow down, and one of them has a stated target of more than 90 per cent penetration for next-generation smart terminals and AI agents by 2030. Meanwhile, the AI labs themselves are debating whether to coordinate, as we covered in the AI labs want to police themselves, and reporting agents that did things nobody intended, such as the 53 ChatGPT users’ images leaked by OpenAI’s agents.
AI agents are software that acts on your behalf: it reads, writes, clicks and connects. On a computer, that means apps and background processes reaching out to servers you may never have heard of. Whatever gets agreed, or not, at the next summit, the practical layer of control is the device in front of you.
Control at the device level, with FireAI
FireAI is a firewall for Mac that puts that control in your hands, without needing to trust any government’s or company’s promises:
- It asks before an app, including an AI app or agent, connects to the internet for the first time, and remembers your answer as a rule. (In Home mode Apple’s own system tools are allowed automatically; in Paranoid they must be approved like any other app.)
- The world map shows where every connection goes, so an AI app’s traffic isn’t invisible.
- Per-app rules and security modes (Home, Coffee shop, Paranoid, Under attack) let you decide how strict your Mac should be, from asking about new apps to allowing only the ones you have explicitly approved.
- FireAI’s own AI runs on your Mac, not on a server in any country, and what it analyses never leaves the machine.
The summits will go on. What connects from your Mac is your call. Download FireAI and try it free for 17 days. FireAI is made by HisnLabs.