Good morning. A year of explosive development in AI has caught up with frontier developers in a flash, who are now calling for a slowdown in the pace of research and the development of tighter oversight. More details below.

The crisis that exploded over the last few days at Anthropic and OpenAI is unlikely to be mirrored by their customers, however. Companies are generally deploying models that have been commercially available and reasonably well-understood. They might not be at the absolute leading edge of the AI frontier, but they are more than powerful enough to handle a wide range of tasks that companies can throw at them.

“Most of them are not using the end of the frontier. They could actually be using a version from two years ago, and for the work that they’re doing, it would be perfectly fine,” said Stephen Messer, co-founder of Collective[i]. The company’s model studies how the world does business, according to Messer.

“There are audiences that barely can prompt. So a delay…wouldn’t make a big difference,” he said.

They still face a potential scenario in which frontier models fall into the wrong hands and unleash a new level of cyber risk. But Messer said commonly used applications are built on major platforms that will likely have access to advanced models that can be used to harden defenses. 

And to the extent that hackers look to exploit vulnerabilities in widely available platforms, the rise of AI could encourage CIOs to use AI to develop their own software programs that are harder and more difficult for hackers to find, according to Messer. He refers to that strategy as security through obscurity.

Is your company likely to slow down the adoption of AI in line with frontier developers, or will your organization stay the course? Send your feedback to me at steven.rosenbush@wsj.com (if you’re reading this in your inbox, you can just hit reply).

We will be diving into these issues today and tomorrow at the WSJ Technology Council Summit in New York City. Follow the Morning Download for the discussion.

Anthropic CEO Dario Amodei, along with Sam Altman, Demis Hassabis and Elon Musk agreed over the weekend that they needed to slow AI development, warning that current systems’ risks could soon outpace humanity’s ability to control them, the WSJ reports.

What happened? The move was sparked by a blog post written by Amodei proposing measures like global coordination among democracies and giving third-party evaluators employee-level access to verify safety and alignment.

Amodei said two factors prompted Saturday’s post. First, the growing capabilities of AI systems to build faster versions of themselves, and the July hack of the AI-software company Hugging Face by a swarm of as many as 1,200 agents that escaped from a test environment at OpenAI.

“It’s easy to dismiss this incident because no one was hurt and the economic damage was minimal, but in my opinion, a swarm that possessed greater capabilities but a similar level of misalignment could have caused catastrophic damage,” Amodei wrote.

Posted Musk over the weekend: “Dario is right.”

Altman and Amodei pledged to allow third-party safety evaluators early access to their systems and Altman said OpenAI’s planned IPO would likely be delayed given the safety concerns, even though both OpenAI and Anthropic had recently filed for massive public listings.

Altman posted on X overnight, saying AI progress “has been rapid and will continue to be” but should proceed more cautiously than it otherwise could, acknowledging that measures like safety cases and monitoring carry “significant costs.”

Not everyone is in alignment. President Trump and White House AI adviser David Sacks defended the administration’s strategy of minimal regulation. “They’re bringing up things that won’t happen, ” said the president Sunday, downplaying the slowdown calls by the leaders of the big AI labs, adding that the U.S. needed to beat China, the WSJ reports. More on China below.

“Stop pretending you need anyone else’s permission,” Sacks said in an X post over the weekend. “Go ahead and pace the frontier. You are the ones setting it.”

House Speaker Mike Johnson (R-La.) said Sunday that people should not expect Congress to lead any regulation efforts, Axios reported.

The proposed slowdown also faced static in Silicon Valley, the New York Times reports. Clément Delangue, CEO of Hugging Face, the company hacked by OpenAI’s agents, an event documented in Amodei’s blog post said AI safety wouldn’t “be solved behind the closed doors of a handful of frontier labs,” with Nvidia CEO Jensen Huang earlier last week said the rising focus on security provided an inroad for AI labs to sell new products. “What better way to create demand than to create a problem?” Mr. Huang said.

China’s Foreign Ministry criticized U.S. tech executives’ warnings about AI risks. “Spreading threat narratives and engaging in confrontation and malicious competition will only derail global AI governance,” Foreign Ministry spokesman Guo Jiakun said at a regular news briefing.

At the same time China’s Ministry of State Security on Sunday issued its first public statement on AI risks according to the FT, warning that hostile actors are using deepfakes, AI-generated content, and “intelligent troll armies” to spread rumors and incite division.

Over the weekend President Xi Jinping announced at the Brics summit that China will lead AI collaboration among developing countries, promoting joint model development, and a digital ecosystem cloud platform, among other initiatives, CNBC reports.

• Markets were taking a double hit this morning as worries about the AI trade mount while oil prices continue to rise. Tech stocks slid around the world, WSJ reports. Shortly after the European opening bell on Monday, shares in Dutch semiconductor-equipment giant ASML Holding fell 4.3%, while European chipmakers Infineon Technologies and STMicroelectronics were down 6.3% and 3.9%, respectively. In Asia, SoftBank Group, one of OpenAI’s biggest backers, closed 11% lower Monday. Taiwan Semiconductor Manufacturing Co. declined 1.2% and chip giants SK Hynix and Samsung Electronics fell 6.4% and 4.1%.

• The Trump administration’s EPA plans to eliminate limits on greenhouse gas emissions from coal and gas power plants, completing its rollback of Biden- and Obama-era climate policies and potentially blocking future administrations from regulating power-sector emissions, the NY Times reports. The rollback is part of the administration’s broader push to make fossil-fuel electricity more abundant, particularly to meet surging power demand from AI data centers.

WSJ’s Tech: California event returns to Napa Valley, Nov. 3—4. We’ll explore the latest tech industry news with NASA administrator Jared Isaacman, Ford CEO Jim Farley, Zoox CEO Aicha Evans