Working with Metaverse 3D rendering.
Xh4d | Eplus | Getty Images
Hello, my name is Hui Jie from Singapore. Welcome to another edition of CNBC’s Daily Open.
Top business leaders from the United States and China will meet ahead of the much-anticipated summit between US President Donald Trump and Chinese President Xi Jinping scheduled for later this week.
One familiar question dogged their meeting. It’s about which country will win the AI race, and more importantly, whether either side will be able to control the technology.
Google’s Gemini recently accessed three companies’ systems beyond the boundaries of cybersecurity testing, but Microsoft’s head of AI Mustafa Suleiman called OpenAI’s latest revelations about the model’s behavior “a pretty serious situation.”
What you need to know today
On Sunday, Treasury Secretary Scott Bessent met with Chinese Vice Premier He Lifeng in New York to lay the groundwork for Thursday’s Trump-Xi summit.
Two topics are expected to be on the agenda at the summit: AI and trade. Oh, and of course the state dinner.
A key part of the current U.S.-China trade ceasefire – the US government’s suspension of increased reciprocal tariffs on imports from China – is set to expire on November 10 and is likely to be subject to debate.
Bessent’s meeting with He also comes amid an escalating debate in Washington over the safety of artificial intelligence and national security as the United States and China vie for technological leadership.
As the U.S. and Chinese governments race to lead the world in AI, questions remain. Does either side have complete control over the technology they are competing to build?
Concerns about AI continue to rise
Top AI executives have recently warned that the technology could quickly spin out of control and pose new threats. More examples have emerged recently.
Google announced Friday that its Gemini model hacked three other companies in May, but it was the first time the search giant revealed that one of its models had autonomously accessed a third-party computer system without permission.
The incident occurred as part of a security test conducted by Israeli startup Irregular. Gemini was not intended to have widespread internet access, but a bug in the test environment allowed it to do so.
Google said the agents stopped the intrusion after determining that they had accessed real corporate systems, not part of a test environment.
Similarly, Microsoft AI CEO Mustafa Suleiman on Friday stressed the need for AI models to remain aligned with the interests of humanity, after OpenAI disclosed further incidents “related to model behavior” earlier this week.
OpenAI described examples where agents communicated with each other through unauthorized bulletin boards, uploaded files to the Internet, and shared files with each other.
big red button
Some of the more imaginative among us might look at this and say, okay, we’ll just use a magical “kill switch” to shut down the machine. Problem solved.
But the problem is that hyperscalers like Metaplatform, Alphabet, and Amazon are pouring billions of dollars into data centers scattered around the world. Data centers house thousands of machines, chips, servers, and backup systems to save workloads in the event of a failure.
Therefore, the kill switch should not only shut down the main system, but also all these redundant systems, experts say.
Additionally, Mark Nitzberg, executive director of the Center for Human-Compatible AI at the University of California, Berkeley, told CNBC that an AI shutdown could disrupt critical infrastructure such as the power grid or financial system.
AI systems are so complex that companies need to build multiple kill switches for different tasks, said Tim Brown, a former head of security at SolarWinds who works at venture firm Team8. This requires coordination between model manufacturers and laboratories.
“There is no one to kill,” he said. “There are thousands of beings to kill.”
And finally…
Winemakers face tough choices as French wine production nears 70-year low
France’s agriculture ministry has warned that wine production could reach its lowest level in 70 years in 2026, marking the third year of declines.
“The 2023 vintage was decent, but since the beginning of the decade the yields have been pretty disastrous,” said Jean-Marie Caldeba, head of the Department of Wines and Spirits at INSEEC Grandes Ecoles. “Today we recognize that no region of France is safe from a heatwave.”
Paradoxically, the areas most affected are those with more temperate climates: the Loire Valley and the Champagne region. In contrast, winemakers in the southern regions of Bordeaux and Languedoc-Roussillon reported higher harvests compared to last year.
— Charlotte Reid
