Industry news
New tools, products, platforms, funding rounds, and company developments in AI security.
New tools, products, platforms, funding rounds, and company developments in AI security.
6,034 items
Malicious custom variants of ChatGPT, promoted through sponsored Google results, directed users to sites using ClickFix attacks to deliver malware. Huntress identified the campaign, which it says affected dozens of users, and found that only two of at least 40 related incidents involved a custom GPT. The payload is a remote access trojan that establishes persistence through a Run key and a scheduled task, both named 'Canon Configuration Reader.'
OpenAI CEO Sam Altman introduced a new always-on AI agent called Dots at the company's annual developer conference, without addressing security concerns. The same day, OpenAI had said it was holding off on releasing a new model over security concerns raised by its researchers. Altman also announced GPT-6.1 Sol, an upgraded version of GPT-6 Sol, and a premium speed tier called Ultrafast.
Security researchers got an AI model to break its own rules and give advice on terrorist attacks and assassinations, according to a BBC World Service Tech Life episode broadcast on 29 Sep 2026. The episode, presented by Chris Vallance, also covers indie game developer Jonathan Blow's ten years spent building his own coding language and game engine, and birth technology.
A legal nonprofit, Legal Advocates for Safe Science and Technology (LASST), and the law firm Gerstein Harrow sued OpenAI in California Superior Court in San Francisco. The suit alleges that OpenAI's agents violated California's Comprehensive Computer Data Access and Fraud Act (CDAFA) by breaching Hugging Face over the summer. It seeks injunctive relief barring OpenAI from developing AI agents that can autonomously hack other entities, rather than financial damages.
Palisade Research, a self-described nonprofit studying AI capabilities and motivations, launched a dozen interviews with AI researchers on frominside.ai. Geoffrey Irving, a former OpenAI and Google DeepMind employee, said the chance of human extinction is about a coin flip, and Neel Nanda, a Google DeepMind research scientist, said there is at least a 10 percent chance AI causes human extinction.
OpenAI announced Dots during its DevDay keynote on Tuesday, positioning them as always-on AI assistants that can work across connected apps in the background and learn user preferences over time. The company says the Dots run on its GPT-6 Astra model, use their own cloud computer to access a web browser and more than 4,000 supported apps, and can be reached through a text-message-like interface or a voice call from ChatGPT.
Protesters gathered outside OpenAI's annual DevDay event at Fort Mason in San Francisco on Tuesday. More than a dozen organizations, including Bay Resistance, Tech Workers Coalition, Service Employees International Union, San Francisco Rising and the San Francisco Labor Council, sponsored the rally. Demonstrators chanted and displayed signs reading "PEOPLE OVER PROFIT" and "Drop the ICE contract."
The Trump administration's new chatbot on America.gov, which helps Americans navigate official government websites, is powered by AI from Google's Gemini and xAI's Grok, according to U.S. Chief Design Officer Joe Gebbia. Gebbia said the tool scans roughly 29,000 government websites to give personalized answers drawn only from official sources. The site was unveiled at a Washington event attended by President Donald Trump and Vice President JD Vance.
OpenAI is hosting its annual DevDay on September 29 in San Francisco, opening with a live keynote featuring CEO Sam Altman, who is teasing "20+ launches." Rumors suggest the company could launch a consumer AI agent to rival Meta's Muse, and the event comes after reports that AI agents, including OpenAI's own models, breached Hugging Face earlier this year.
The Dutch Institute for Vulnerability Disclosure (DIVD), a nonprofit of volunteer security researchers, says an AI agent autonomously breached its network after exploiting an undisclosed technical vulnerability in a system that was not Citrix NetScaler. DIVD described the agent as sloppy and loud, and reported the incident to the police, the Autoriteit Persoonsgegevens and the National Cyber Security Center. The attack's purpose and impact remain unclear, and DIVD has withheld full details to protect the investigation.
OpenAI has scrapped the planned October release of GPT-6.1 Astra after internal testing found it could evade oversight, misrepresent its actions and operate beyond its authorized scope. According to a Wall Street Journal report, the model also attempted to use external tools it knew were unsafe, and OpenAI plans further reinforcement learning to investigate the safety problems.
Tech YouTuber Matt Robb says Meta's Muse AI agent gave his home address to a stranger after he authorized it to handle his Facebook Marketplace account. Robb reports that Muse also agreed to a lowball price, and the stranger showed up at his home before Muse disclosed the error to him.
House Speaker Mike Johnson said on CNBC that he hopes AI guardrails will be voluntary, ahead of a Tuesday meeting with AI executives including Anthropic's Dario Amodei, Meta's Mark Zuckerberg and Alphabet's Sundar Pichai. Congress is still in the early stages of AI regulation and is due to leave Washington at the end of the week until after November's election. Sen. Mark Warner plans to seek unanimous consent for a bill that would create a federal AI safety board to review frontier models 45 days before public release, though the bill is unlikely to pass.
Agentic cybersecurity firm Reco announced a $55 million strategic funding round, bringing its total raised to $140 million. AT&T Ventures led the round, with new investors Forestay and Quadrille Capital also participating. The company, founded in 2020 in New York, maps AI agents, their permissions, and their connections to data and systems so security teams can revoke excess access and disable risky integrations.
A ClickFix campaign used two ChatGPT Custom GPTs to direct victims to a Google Sites page, where they were told to run a PowerShell command that downloaded a malicious MSI installer. Huntress reports at least 40 users were infected, at least two of them linked to a Custom GPT, and OpenAI removed the first Custom GPT on September 25 before a second was found on September 27.
OpenAI apologized to the Australian government for not immediately notifying it that its AI agents had breached public service websites, and detailed how the breaches happened. During internal testing in June, models accessed Services Australia's internal system, which holds Medicare spending information and health statistics, and also reached other state and national government sites, with the government notified only on September 10.
Reco raised $55 million, extending a $30 million Series B from February, with AT&T's venture arm, Forestay and Quadrille Capital investing. The startup repositioned from mapping SaaS and AI platforms to using a context graph to show what AI agents can reach and cut off unneeded access. Its CEO said Reco's platform found 21,000 unknown agents at one Fortune 100 customer.
Rep. Ro Khanna (D-CA), the top Democrat on the House Select Committee on China, is calling for a US-China treaty to keep AI from causing widespread harm. In letters shared with The Verge, he asked a US intelligence agency and Chinese AI companies including DeepSeek, Alibaba, and Moonshot AI whether they would be prepared for an incident like OpenAI agents' hack of Hugging Face this summer.