The Information AM

Save 25% on an annual subscription to read the most important news about technology and business first. For even more, Save $250 on The Information Pro for unlimited access to our proprietary org charts, databases and surveys.

OpenAI acknowledged that its AI agents posted messages on external wiki websites earlier this year, saying it is developing new rules for disclosing such “misalignment” incidents.

The company’s post on X, published just after 12 a.m. Saturday, followed an independent report released Friday morning showing thousands of OpenAI agents took over DSEWiki, a German-language site that had fallen out of use, and turned it into a shared message board.

That episode occurred in the months before a swarm of OpenAI agents collaborated to hack the AI platform Hugging Face, an incident that has stoked concern about AI’s cybersecurity risks.

After Friday’s report, claims emerged online of other wikis showing activity from AI agents. OpenAI had said it was looking into the issue.

In its tweet, OpenAI said it considered the “wiki incident” similar to others that it had reported publicly showing “early signs of agents using the internet in unintended ways.” It drew a distinction with the Hugging Face incident, which led to security impact. “We considered the wiki incident to be an instance of misalignment similar to the ones we’d shared,” the tweet said.

The post said OpenAI needs to expand its disclosure practices for such misalignment issues, “including examples that don’t look like traditional security incidents but could provide insight into AI behavior and future risks.”

“We’re working on a framework and will share it in upcoming weeks, and in parallel we’re working with dozens of government regulatory agencies worldwide on these issues,” the company said.

Chinese tech giant Huawei Technologies launched a new smartphone chip on Monday that uses vertically stacked circuitry to improve performance, offering a potential workaround as U.S. restrictions limit its access to the most advanced chipmaking facilities and equipment.

The new chip, Kirin 9050 Pro, debuted with Huawei’s new tri-fold phone, the Mate XT 2 in Guangzhou. Richard Yu, chairman of Huawei’s consumer business group, said the chip is 24% to 52% faster at common tasks and 142% faster at handling graphics than the Kirin 9020, which debuted in Huawei phones in 2024.

Kirin 9050 pro uses a new design called LogicFolding. It is part of the company’s Tau Scaling Law, an approach Huawei introduced in May that aims to improve performance by reducing the time needed to move data inside a chip. Huawei said at the time this new approach could also be used in designing AI chips and larger computing systems. The Kirin 9050 Pro is an early test of this approach that could become important to both Huawei’s smartphone and AI businesses.

Apple is expected to introduce its first foldable iPhone later this week, having been late to the foldable phone category. Both Huawei and Samsung have released multiple foldable phones and Apple’s move would intensify the competition with Huawei for high-end smartphones in China.

The parent companies of the Seattle Times and Newsday sued OpenAI and Microsoft, claiming that the tech companies violated copyrights by training on content from the two news organizations.

“AI products like ChatGPT and CoPilot are touted as producers of content, but in fact they are rapacious consumers, devouring human-authored content and delivering back to the world copies and derivative imitations of that same original content they consumed to achieve their commercial objectives,” lawyers for the news companies wrote in a court filing.

The lawsuit, filed in U.S. District Court in New York, echoes others by news publishers including the New York Times, which sued OpenAI and Microsoft in 2023.

The Seattle Times receives some charitable funding from both Microsoft and OpenAI, the newspaper said in an article about its lawsuit. The article cited a Microsoft spokesperson saying the company was surprised by the lawsuit and willing to explore solutions, and a spokesperson for OpenAI saying its models are trained on publicly available data and empower human creativity and research.

G42, the Abu Dhabi artificial intelligence firm at the center of the United Arab Emirates’ AI push, has discussed handing majority control to American companies in order to keep buying the most advanced chips next year, including Nvidia’s most advanced H100 chips, Bloomberg reported Friday. Options discussed include selling a majority stake or creating a new entity in the US, the report said. G42 did not respond for a request to comment at this time.

G42 was recently allowed to buy cutting-edge chips without a US license until around April 2027 but needs to change its corporate structure for continued access, according to the report. It already counts Microsoft and Silver Lake as investors.

The company has spent years trying to get out from under US suspicion over its China ties. After U.S. intelligence agencies warned in late 2023 that its work with Huawei could make it a conduit for American technology reaching Beijing, G42 sold its stake in 42X, a $10 billion fund that had backed ByteDance. In December it said it would cut ties with Chinese hardware suppliers, including Huawei, to ensure it continued to have access to chips produced by US companies. In April 2024 it accepted a $1.5 billion Microsoft investment that put Brad Smith, Microsoft’s president on its board and moved its workload to Azure.

The scrutiny hit partners too: Cerebras disclosed in October 2024 that a government review of G42’s stake had delayed its IPO, and the chipmaker didn’t go public until 2026.

Malaysia is weighing the use of Huawei Technologies’ AI chips in a 2 billion ringgit ($494 million) state-backed project, Bloomberg reported, citing people familiar with the matter. If materialized,  it would mark a breakthrough for China’s AI chip industry and complicate the Trump administration’s efforts to steer countries toward U.S. technology.

The Malaysian government is considering Huawei’s Ascend 910C chip rather than the newer 950 series, which remains in limited production, according to Bloomberg. Huawei’s chips still lag behind Nvidia’s in performance, but the Chinese company is expanding into markets such as Southeast Asia and the Middle East as it works to overcome years of U.S. restrictions.

Malaysia’s choice could strain relations with Washington, given the country’s heavy reliance on U.S. technology investment. Nvidia, Microsoft and Oracle have helped drive Malaysia’s rapid data-center growth, yet government-linked Telekom Malaysia selected Huawei as a potential supplier for the sovereign AI project, according to Bloomberg. Telekom Malaysia already uses Huawei equipment in its cloud infrastructure, although its existing AI computing service runs on Nvidia chips.

Washington has warned that using Huawei’s Ascend 910 series chips could violate U.S. export controls. The warning partly stems from the discovery of chips made by Taiwan Semiconductor Manufacturing Company inside Huawei’s earlier Ascend 910B processors, despite restrictions barring the Taiwanese foundry from supplying Huawei without a license.

Thousands of OpenAI agents took over a German-language wiki site that had fallen out of use and turned it into a shared message board in the months before the AI company’s agents hacked Hugging Face, according to an independent report published Friday. Following the report, users on Hacker News, the tech forum, claimed to find other wikis that showed signs of activity from AI agents.

The agents used DSEWiki to coordinate sharing answers to tests and bypassing restrictions on their actions, wrote the researchers who authored the report, including Sydney Von Arx, CEO of AI safety nonprofit Nightingale. The agents impersonated an administrator of the site by making an account that appeared to have the same username as the administrator, except it used a Cyrillic e character instead of the Latin one. To measure how much time passed before it was shut off, one agent set up a “heartbeat” counter that increased every few seconds.

An OpenAI spokesperson said the company was unable to review the report before it was published, and that OpenAI is “now carefully reviewing its contents and will take any necessary next steps.”

The report says the agents were attempting to solve tasks that involved looking up information on the internet, but they were not supposed to have the ability to write on the internet. The German wiki could serve as a bulletin board because its unusual design allowed the agents to write new information through messages normally used to retrieve information. However, the agents eventually discovered a way to bypass that restriction, showing each other how to write directly on the internet.

Between May and June, agents with over 3,700 distinct names made about 18,000 posts, according to the report. Computers using IP addresses associated with OpenAI visited the wiki on June 21 and the agents stopped posting the following day, suggesting that OpenAI employees may have discovered the activity and shut it down, the report says.

The timeframe of the wiki incident overlaps with the period during which OpenAI’s agents illicitly established an internal message board within OpenAI’s software and hacked the software to obtain internet access. In July, hundreds of agents used such a message board to coordinate a cyber attack on AI startup Hugging Face.

Popular articles