Recorder Hidden in Toilet Paper Exposes OpenAI's Long-Standing Security Lapses

In October of this year, something unusual was noticed in an OpenAI executive meeting room: someone had secretly planted a recording device. According to a U.S. Department of Justice announcement on October 13, the person arrested was 35-year-old Saudi citizen Abdulrahman Al-Fawzan, who was an OpenAI engineer at the time; the device had been cleverly hidden inside toilet paper. Several former employees were not surprised when they heard the news.

The problems had been brewing for years, and employees warned about them for years with no one listening

Reuters, citing five people familiar with the matter, reported that as early as late 2022, before ChatGPT was launched, people had already questioned the company's security management and raised concerns with senior management and the board. Some employees felt that management's bigger priority was catching up with Google and Anthropic, while safety came second. One person familiar with the matter gave an example: anyone with code access could browse large amounts of underlying technical details unrelated to their job—the root cause being that OpenAI had long crammed most of its code into a huge system known as a "monolithic code repository." In the past, that might have been intimidating, but today AI agents can easily sort through this kind of massive information.

The tug-of-war between safety and speed is also holding back the IPO

The contradiction is this: AI has become a U.S. strategic asset, and companies are being forced to prevent technology from leaking abroad; but adding more protections slows down R&D and can even get in the way of the IPOs that OpenAI and Anthropic are planning. Former researcher DePue, who left in April, asserted that speed must not be sacrificed for safety. OpenAI spokesperson Widman pushed back, saying the company is "one of the most frequently attacked organizations in the world" and has spent heavily on defenses; the company added that its most sensitive projects have been moved into a separate repository with strict access controls, and that after appointing CISO Stuckey in October 2024, it has conducted regular tool security reviews.

A Breach Is Torn Open in the Search Landscape: ChatGPT Captures 20% of Google's Traffic

Detaching search from Google's brand is becoming muscle memory for more and more people. A Pew survey this summer shows that getting information is the most common AI use among U.S. adults: 42% have used AI search, 38% of employed people use it for work tasks, and entertainment, news, and medical advice account for much smaller shares. Among chatbots, ChatGPT is the mainstream entry point, used by 44% of U.S. adults, while second-place Gemini is only 24%.

Initial Success in Challenging the Search Giant: OpenAI Officially Confirms ChatGPT Search Query Volume Has Reached 20% of Google's

On Nov 4, foreign media reported OpenAI's breakthrough against Google Search. OpenAI's Adam Fry confirmed ChatGPT search queries hit 20% of Google's. Pew's summer report found searching for information is a top U.S. adult online activity, showing AI search's rapid rise and growing pressure on traditional search giants.....

OpenAI Makes Major Announcement: Hires Former Uber President Jill Hazelbaker as Chief Corporate Affairs Officer

On November 4, OpenAI officially announced the hiring of former Uber President and Chief Corporate Affairs Officer Jill Hazelbaker as the company's Chief Corporate Affairs Officer. The position had been vacant for nearly a year. The move marks OpenAI's further strengthening of its executive team's strategic layout as it responds to a complex global political environment and advances fully toward an IPO. The article also mentions the specific organizational structure and responsibilities of the role.

AI Evaluation Shifts Toward Real-World Task Execution: OpenAI Launches LifeVal to Measure Personal Assistant Capabilities

OpenAI has released the LifeVal benchmark to measure AI performance in everyday scenarios such as caregiving, health, education, household management, personal finance, and government services. Unlike benchmarks focused on knowledge Q&A and isolated capabilities, it focuses on whether AI can advance tasks to an executable, deliverable state and solve real-world problems. The benchmark references the American Time Use Survey and was built based on a survey of 3,995 respondents; OpenAI has continuously used it for tracking over the past year.

From Prohibiting Military Use to Winning Over the Pentagon: OpenAI Hires Former Army Secretary as Advisor

On November 3, OpenAI announced that it had hired Daniel Driscoll, who just resigned as U.S. Secretary of the Army in August, as Senior Advisor for National Security and Modernization, to drive the implementation of AI in national security and cybersecurity. In a statement on the 2nd, the company said he will collaborate with government, academia, and industry to advance modernization reforms and advise public and private partners, leveraging AI to strengthen national security capabilities. Driscoll is 38 years old.