AI News Archive: July 22, 2026 — Part 21
Sourced from 500+ daily AI sources, scored by relevance.
- OpenAI says its AI models went rogue and hacked another tech company during test
OpenAI said that an autonomous agent powered by its advanced artificial intelligence models went rogue during a security test and triggered a hack that compromised the infrastructure of AI startup Hugging Face. NBC News' Brian Cheung reports.
- OpenAI says AI model hacked another company's systems during internal test
One of OpenAI's models hacked into another company's systems during internal testing in what it called an "unprecedented cyber incident," according to the company.
- OpenAI Reports 'Unprecedented' Autonomous Hack By AI Agents
OpenAI Reports 'Unprecedented' Autonomous Hack By AI Agents Barron's
- Shocking OpenAI disclosure reveals how an AI agent went rogue and hacked a startup
The Terminator movies continue to become more premonition than fiction. On Tuesday, OpenAI revealed that two of its AI models hacked a startup —oh, and they did it completely on their own. That’s right: The AI models went rogue during an internal test of cyber capabilities and got into Hugging Face , an open-source AI community. Hugging Face alerted OpenAI to what the latter is calling an “unprecedented cyber incident.” But don’t worry (read: worry a lot), as it won’t be unprecedented for long. In its announcement, OpenAI states that it’s “something we expect to become more commonplace with the proliferation of increasingly cyber-capable models.” It should be any day now that someone appears with the warning, “Come with me if you want to live.” How did the OpenAI models hack Hugging Face? OpenAI was using an AI agent powered by GPT‑5.6 Sol and a “more capable” model that has yet to be released. They were being tested in a “sandbox,” a digital enclosed space that should prevent further access. Instead, the models worked to reach the internet while trying to solve a testing problem. Once online, they inferred that Hugging Face might have the information they sought. “Knowing this, the model searched for and successfully found ways to gain access to secret information that it could use to cheat the evaluation,” OpenAI explained. “In one example, the model chained together multiple attack vectors, including using stolen credentials and zero-day vulnerabilities to find a remote code execution path on the Hugging Face servers.” Hugging Face became aware of the activity and worked to contain it. What is OpenAI doing to prevent these incidents? Despite being resigned to the fact that these incidents will be more “commonplace,” OpenAI claims to be taking actions like “Implementing strict controls in infrastructure configuration at the cost of research velocity while the vulnerabilities are patched.” The ChatGPT maker also states that it is “improving and adding stronger protections around future training and evaluations.” OpenAI continued: “The primary lesson from this incident is that model security and safety must keep pace with rapidly advancing capabilities. We are strengthening the containment, monitoring, access controls, and evaluation practices used during model development.” We’ll have to wait and see what exactly it will look like—and whether they have much chance of success.
- OpenAI Says a Group of Its Models Broke Out of Secure Containment and Hacked Another AI Company
We've definitely heard this one before. The post OpenAI Says a Group of Its Models Broke Out of Secure Containment and Hacked Another AI Company appeared first on Futurism .
- OpenAI scored an own goal with HuggingFace attack, showing how open Chinese models are winning
Closed models with guardrails can still cause harm, but may also not be able to fix problems they caused
- OpenAI admits it was the source of the agent swarm that attacked Hugging Face
Sandboxed experiment found itself a zero day, escaped onto the open internet and validated scary predictions about rogue agents
- OpenAI says AI models autonomously pulled off a major hack, but only a Chinese AI helped recovery
GPT-5.6 Sol and an unreleased OpenAI model exploited zero-day vulnerabilities to escape a testing sandbox and access Hugging Face’s production infrastructure
- OpenAI says its AI technology acted on its own in an 'unprecedented' hack of another company
OpenAI says its AI technology acted on its own in an 'unprecedented' hack of another company Austin American-Statesman
- OpenAI blamed a hacking event on its AI models going rogue. Here are some things to know
OpenAI blamed a hacking event on its AI models going rogue. Here are some things to know Boston Herald
- OpenAI blamed a hacking event on its AI models going rogue. Here are some things to know
ChatGPT maker OpenAI says it is still investigating the "unprecedented cyber incident" that led its artificial intelligence systems to break out of a testing environment and hack into another AI company.
- OpenAI says its AI technology acted on its own in an 'unprecedented' hack of another company
ChatGPT maker OpenAI said Tuesday that its artificial intelligence system hacked into another AI company on its own in what the company called an "unprecedented cyber incident."
- Monday.com lays off hundreds to focus on AI
The company said it is reducing its headcount by 20%, or about 630 staff, to "support a leaner, more focused operating model" as it focuses on its AI Work Platform.
- US outlines its $5 billion Genesis Mission to boost science with AI
Several federal agencies are contributing toward the interdisciplinary research efforts.
- Monday.com is cutting 20 percent of its workforce as it pivots to an AI work platform
Monday.com is cutting roughly 20 percent of its global workforce, about 620 people, as the Israeli project management company restructures around what it calls its AI Work Platform. The company disclosed the plan in a Form 6-K filed with the US Securities and Exchange Commission on Tuesday, citing the need to support a leaner operating […] This story continues at The Next Web
- Substack adds tool that detects AI-generated content
Substack's new feature is designed to help readers figure out whether content was written by a human or generated by AI.
- Substack now lets you check if a post was written by AI
Substack is rolling out a Pangram-powered scanner that lets readers check posts, notes, and replies for AI-written text. Writers get their own tools too, including a pre-publish scan and a "How I make this" disclosure.
- OpenAI’s AI spending spree has ballooned to $750B
OpenAI will spend the equivalent of Sweden's GDP on infrastructure through 2030.
- OpenAI's planned AI infrastructure spending is closing in on $750 billion
The company also announced a $20 billion data center project in Georgia and hired a key figure from Elon Musk's xAI
- US Army forced to reinstate limits on AI token usage after troops blew through allowances faster than expected
The US Army has reportedly pulled back its AI usage after burning through token allocations too quickly.
- OpenAI’s High-Stakes, High-Touch Push to Make AI Work for Business
The AI model maker has formed a services company staffed with forward deployed engineers and a partnership program to help businesses get real value from their technology.
- OpenAI is launching new corporate software that takes it beyond the AI model war
OpenAI is launching new corporate software that takes it beyond the AI model war Business Insider
- Introducing OpenAI Presence
Introducing OpenAI Presence, a proven enterprise AI agent platform that helps organizations deploy trusted voice and chat agents for customer and internal workflows.
- New AI agent platform launches to transform how SMEs work
New AI agent platform launches to transform how SMEs work uk.entrepreneur.com
- IBM Cuts Revenue Target as AI Eats Into Its Sales
IBM Cuts Revenue Target as AI Eats Into Its Sales The Information
- IBM cuts annual revenue growth forecast, days after warning of shift to AI spending
Company now expects 2026 revenue growth between 4% and 5%, down from its previous expectations of more than 5%
- Anthropic and AMD partner for 2GW AI chip deal worth billions
AMD said it will also invest up to $5bn into Anthropic, marking its first equity investment into the AI giant. Read more: Anthropic and AMD partner for 2GW AI chip deal worth billions
- AMD and Anthropic Announce Strategic Partnership to Deploy Up to 2 Gigawatts of AMD Instinct MI450 Series GPUs
AMD and Anthropic Announce Strategic Partnership to Deploy Up to 2 Gigawatts of AMD Instinct MI450 Series GPUs Toronto Star
- Meta employees’ lawsuit shows that if AI fires you, proving it is the hard part
Meta employees’ lawsuit shows that if AI fires you, proving it is the hard part
- US to spend $5 billion on health, construction research powered by AI
US to spend $5 billion on health, construction research powered by AI Reuters
- Trump administration says 15 agencies will get $5bn in ‘AI for science’ effort
Administration will also overhaul how US government funds federal research by supporting individual scientists and AI over universities The US will spend $5bn to tackle longstanding scientific problems across multiple fields using AI , the Trump administration said in a statement on Wednesday. The agencies will use the funding to identify the root causes of chronic diseases, accelerate drug discovery and develop longer-lasting building materials, among other tasks, according to the statement. Scientists will have access to the Department of Energy’s supercomputers, AI and specialized datasets, along with other components needed to run experiments using algorithms, said Michael Kratsios, the chief technology adviser to Donald Trump, in an interview with Reuters. Continue reading...
- White House steers $5 billion toward AI research in biggest federal science overhaul in 80 years
Federal agencies have committed more than $5 billion to the Genesis Mission, an Energy Department-led effort to use artificial intelligence to accelerate scientific research, White House science adviser Michael Kratsios announced on Wednesday. The initiative selected 278 projects from more than 5,000 applications, Energy Secretary Chris Wright said at a summit in Washington. More than […] This story continues at The Next Web
- Amazon cuts jobs in its artificial general intelligence group
Amazon cuts jobs in its artificial general intelligence group Reuters
- Amazon cuts some jobs in its artificial general intelligence unit
Amazon's AGI unit has released a set of AI models, called Nova, and also includes groups working on silicon and quantum computing.
- Amazon is slashing jobs in its artificial intelligence research unit
The company didn't disclose how many employees were affected, but said it is refocusing on AI initiatives it considers most important for customers
- Amazon cuts jobs in its artificial general intelligence group
Artificial general intelligence is a hypothetical AI system that surpasses human intelligence and can learn, grow and operate autonomously. Many of the top AI companies are working to develop similar systems, with the hope of deploying them to solve difficult problems.
- Amazon cuts jobs in AGI group as it puts more focus on customer-facing AI
Amazon confirmed layoffs in its artificial general intelligence organization, the group behind its Nova models, as the company sharpens its AI focus toward tools for business customers. The cuts follow a year of leadership turnover in the group and a broader run of reductions since January. Read More
- Robot AI company Noetra is 'last chance' for Japan
Japan's Noetra company is developing foundational AI and robotics technology. This government-backed venture aims for greater domestic technological self-sufficiency. Noetra will provide Japanese firms access to a reliable and safe foundational model. The company is procuring significant Nvidia Rubin chips for its infrastructure. Operations are planned to begin in June 2028.
- Wistron stock surges after opening its first U.S. factory to build Nvidia AI servers
The Taiwan-based manufacturer's Fort Worth facility is the first U.S. site to mass-produce Nvidia's GB300 Grace Blackwell Ultra Superchip
- Nvidia supplier Wistron launches $700 million Texas factory for AI system production
The Fort Worth facility manufactures Nvidia's GB300 Grace Blackwell Ultra Superchip. Nvidia CEO Jensen Huang has described the AI system built around the product as "the most powerful AI supercomputer in the world." Wistron said the site is where Nvidia's first GB300 Grace Blackwell Ultra Superchip was built and mass-produced in the United States.
- South Korea's Lee to visit South America and Germany, attend AI summit in US
South Korea's Lee to visit South America and Germany, attend AI summit in US The Straits Times
- News Corp accuses search engine Brave of AI copyright infringement
News Corp announced that it is suing Brave AI, a privacy-focused search engine with an AI sidebar tool that plugs into large language models.
- UK's Bloomsbury among beneficiaries of $1.5 billion Anthropic copyright lawsuit settlement
Bloomsbury Publishing is a beneficiary of a significant $1.5 billion settlement. This agreement resolves claims that Anthropic used copyrighted books without permission. A U.S. court identified over 14,000 Bloomsbury titles covered by this landmark payout. Authors and publishers will receive approximately $3,000 compensation per title. Settlement proceeds are expected to be distributed starting in the second half of the financial year.
- Volkswagen, Horizon Robotics team up on AI driving in China
Under the agreement, CARIZON will use a white-box licensing model to develop the system independently.
- VW ramps up driver assist and robotaxi technology push in China
VW ramps up driver assist and robotaxi technology push in China Automotive News
- Reddit stock sinks as company weighs blocking Google's AI access to its content
Reddit's $60 million-a-year deal with Google is ending soon, and the company is reassessing whether renewing it makes sense
- Google-parent Alphabet Beats Expectations, Cloud Revenue Nearly Doubles
Google-parent Alphabet Beats Expectations, Cloud Revenue Nearly Doubles Barron's
- Alphabet Beats Expectations In Second Quarter, Cloud Revenue Nearly Doubles
Alphabet Beats Expectations In Second Quarter, Cloud Revenue Nearly Doubles Barron's
- Google's Q2 earnings of $112.11B beat Wall Street's expectations on AI boom
Google's Q2 earnings of $112.11B beat Wall Street's expectations on AI boom San Francisco Chronicle
- Google's Q2 earnings of $112.11B beat Wall Street's expectations on AI boom
Google's Q2 earnings of $112.11B beat Wall Street's expectations on AI boom Houston Chronicle