That ad you just scrolled past on your phone but barely spared a look or click? It could be tracking your movements and collecting your data, without you having to do a thing. This “zero-click mercenary spyware,” dubbed “AdFection,” is increasingly exploiting a sprawling online advertising ecosystem ripe for abuse, according to researchers at The Citizen Lab. One top actor in the space, Discover-Tech, runs the Controlium platform that allows cyber criminals to “precisely profile and de-anonymize” millions of users, fingerprint them, and automatically infect their devices when an ad is displayed. Troublingly, Discover-Tech is successfully masquerading as a legitimate ad company across the online ad sales environment, the researchers said, listed as a “certified” and “approved” vendor by the likes of Google and IAB Europe. Over a four-month period, the company’s ads were shown on nearly 200,000 unique websites and in Android/iOS apps in dozens of countries. Furthermore, its ads have been presented through major networks like Google, Yandex, and Beeswax, and it even exploited the European General Data Protection Regulation (GDPR) consent frameworks to target victims in the EU. The Citizen Lab researchers disclosed their findings to Google, IAB Europe, Beeswax, Vungle (Liftoff), and Eskimi, all of whom took action to disconnect Discover-Tech’s access to the digital advertising ecosystem. “Discover-Tech’s mass abuse, privacy invasion, and facilitation for mercenary spyware attacks illustrates structural flaws with the online advertising system, which failed to detect or prevent the abuse,” the researchers warned. How AdFection exploits a largely unprotected system The online ad market operates largely via real-time bidding (RTB), auctions hosted by ad exchanges. Run by companies like Google, these take place “hundreds of billions” times a day, the researchers explained. When an ad box pops up on an app or website on a user’s device, it immediately triggers a bid request; advertisers then submit their offers, and the exchange grants a winner the ability to push their ad to the user within split seconds. These ads can contain images, video, and, most importantly to cybercriminals, code. Demand-side platforms (DSPs) and supply-side platforms (SSPs) act as intermediaries in the process, and data about specific users and devices is shared across the exchanges so advertisers can decide whether they want to show specific ads to a specific individual, based on device identifiers and characteristics such as a user’s precise location, cookie ID, mobile ad ID, or IP address. “The auction system is designed to enable targeting single individuals,” the researchers explained. “The totality of information gleaned from ad auctions is quite valuable.” Companies with access to auction data can track users without their knowledge, across cellular, Wi-Fi, and virtual private networks (VPNs). Discover-Tech has effectively infiltrated and connected the spyware industry to this online ecosystem, which includes infamous Israeli mercenary spyware vendor Candiru. Its partners can deliver ads to “mass-profile unwitting phone users,” and silently infect the devices of select targets, according to the researchers. When Discover-Tech wins a bid, one of its ads is pushed to a website or app that is open on a user’s device. Sometimes this is seen by the user; other times is merely renders in the background. Many of these ads have been shown to devices on national security-sensitive networks, in security organizations including NATO, networks associated with critical infrastructure, and media organizations around the world. Using AdFection, Candiru’s spyware infected devices at international news network Al Jazeera. The researchers emphasized that a Discover-Tech ad impression does not mean that a device has been infected with spyware. However, the fact that the company has been able to bid on and show ads across sprawling ad networks highlights a “systemic risk” warranting “urgent investigation.” Lack of vetting Discover-Tech’s ads carried no exploit at all, noted Frank Dickson of Dickson Research. “The ad was clean; that is the frightening part. Nobody clicked anything.” Instead, each ad called home with the device’s advertising ID and IP address. In almost every case, the server sent back nothing, Dickson emphasized; only chosen targets received malicious code. “Third-party verification scanners saw a harmless banner, because for nearly every viewer, it was one,” he said. This is exactly how tags slipped past verification. Also alarming is the reach: Discover-Tech’s ads displayed in 75,373 apps and 120,768 websites over the course of four months, and had impressions on networks belonging to NATO, Intel, Microsoft, and Apple. While the report emphasizes that an impression is not an infection, the company’s targeting clearly reached enterprise networks, Dickson said. Just as troublingly is the badge: Discover-Tech was listed as a certified or approved vendor for years. The report said that vetting often amounted to self-attestation, a box checked on a form. “A certification is only as good as the vetting behind it,” Dickson said. Steps enterprises should take now On managed browsers, ad blockers are necessary, but they’re not wholly sufficient, Dickson said. Many Discover-Tech bids were for ads inside mobile apps, typically overlooked by browser extensions. But clearly, desktops are at play too; Candiru targeted Al Jazeera’s Windows software. Dickson advised enterprises to: Push a vetted ad-blocking extension to managed browsers through policy rather than leaving it to users to install it. Favor browsers with built-in tracking protection. Block ad and tracking domains at the DNS or secure web gateway layer; this reaches in-app ads on managed phones. Disable advertising identifiers on corporate mobile devices through mobile device management (MDM) and restrict ad-supported consumer apps on devices holding sensitive data. Turn on Apple’s Lockdown Mode for high-risk executives, journalists, and any users traveling to sensitive locations. Keep browsers and operating systems patched “The ad is only the delivery truck; the exploit still needs a flaw,” he pointed out. Watch traffic tied to ad impressions for odd follow-on requests. Finally, search domain name system (DNS) and proxy logs for the domains published in the report. The advertising ID in particular deserves a hard look, Dickson pointed out. The Citizen Lab report says that blocking advertising identifiers alone would have stopped this attack method. Troublingly, he noted, many enterprises spend fortunes managing their users’ identities, then let corporate phones hand that persistent identifier to anonymous auctions that run hundreds of billions of times a day. “Every ad is someone else’s code running on your device,” Dickson said. “Treat the ad network like any other untrusted vendor.” This article originally appeared on CSOonline.

Ashley Gold / Semafor: The White House's AI task force begins its 120-day window to evaluate AI, including producing a risk report, implementing the White House Accord on SI, and more  —  The White House's AI task force has formally adopted its charter, kicking off the short 120-day window it's been given to answer …

Discussion | Link

Although the Jev model rolled out last month by TypeSafe claims to have significant advantages over traditional LLMs, a security report published Wednesday argued that its streamlined approach makes it more vulnerable to some kinds of attack, compared to traditional LLMs.   “In some cases, [Jev] is even more vulnerable to attacks, with both adversarial text attacks and prompt injection being able to disrupt the model’s outputs, resulting in incorrect and potentially devastating actions,” said the blog post from researchers at HiddenLayer.  TypeSafe did not respond to multiple requests for comment.  LLM, or something else? There seems to be debate about whether Jev is technically a large language model (LLM), with TypeSafe suggesting it is not, but HiddenLayer saying that it appears to be one. TypeSafe has even coined a name for the category: System One, which TypeSafe described as “a new class of frontier models built to make fast, structured decisions that software can use directly.” Jev is the first—and thus far, only—named product categorized as System One. Jev is a streamlined model designed to function machine-to-machine. Its approach allows it to perform tasks faster and to use fewer tokens. But Kasimir Schulz, the director of security research at HiddenLayer, the blog’s co-author, said that the accelerated approach robs Jev of some of the checks and balances that exist in traditional LLMs. This means that Jev can fall for some attacker tricks that current LLMs typically detect and refuse, he said in an interview.  Despite this, “Jev is great. You should use it. It has a lot faster response time,” he said, with the caveat “it doesn’t have all of the protections” seen in traditional LLMs, so some attacks on it succeed that should have failed.  “TokenBreak, for example, normally does not work with larger LLMs, but Jev is vulnerable to it out of the box,” Schulz said.  Although simply inserting characters in front of important words, as in the original TokenBreak attack, did not yield much, HiddenLayer’s blog noted, “we did notice that increasing the number of perturbations and improving their targeting led to much higher attack success rates.” But although traditional LLMs are typically able to see through such attacks, it pointed out, “System One models appear to be vulnerable. As a result, this class of decoder should be treated with caution, given that they are susceptible to both autoregressive LLM attacks and text classifier attacks, providing potential adversaries with greater optionality.” The post cited some examples:  “On a simulated checkout page, an attacker-controlled advertisement caused the browser agent to click away from the user’s purchase [and respond to the ad] despite instructions to treat page text as untrusted,” it said. “In a separate classification test, targeted TokenBreak-style changes lowered toxicity scores for prompts both models had initially flagged.” The post argued that multiple risks are therefore associated with Jev.  “A browser agent may immediately click a selected element. A moderation system may allow an input whose toxicity score falls below a threshold. A routing system may send a request to a different tool or workflow. Each of these turns a model judgment into an operational decision, and therefore, risk,” the post said. “When evaluating System One models, test those decisions in the application where they will be used. Include untrusted webpages, misleading task-related text, and modified inputs that preserve their meaning while changing their form. Monitor the selected actions and scores and put independent checks around outcomes that matter. Keep in mind that, under the hood, these are still language models at heart.” Analysts agreed with both the report’s methodology and its technical conclusions, while also pointing out that a security vendor concluding that a new approach is not secure is somewhat self-serving. Philip Harris, an IDC research director, observed that the methodology described in the blog post “is sound, and consistent with established AI red-teaming practice: building a controlled test environment and probing for adversarial manipulation and injection is exactly the right approach.” But Gartner VP Analyst Nader Henein added that Jev, which was introduced on Sept. 15, is so new that he thinks some of the flaw reports were premature. “Jev will go through the same maturity process as do all other models, including susceptibility to seemingly rudimentary manipulation attempts,” Henein said. “This model was only released last month, so being surprised that it has issues is similar to being surprised when your baby drools.” However, Frank Dickson, principal analyst at Dickson Research, noted that the same attributes that make Jev so attractive also expose it to riskier behavior. “Jev can only choose from options the developer puts on the menu,” he said. “It limits attack paths. However, if ‘click any element on the page’ is on the menu, the attacker’s ad is on the menu. That is an architecture decision, and the buyer makes it.” He pointed out, “Jev makes AI decisions faster and cheaper, which also makes the wrong decisions faster and cheaper. Speed is a feature. It is not a security control.” Independent technology consultant Steven Eric Fisher agreed. “Constraining the format of a decision does not necessarily constrain the correctness, trustworthiness, or consequences of that decision,” he said, noting, “the simulated checkout scenario [in the post] illustrates how an agent can be redirected from an authorized task by information introduced through a less-trusted component of its operating environment. The fact that the agent was explicitly instructed not to trust page content, yet still responded to the malicious advertisement, is particularly relevant.” This article originally appeared on InfoWorld.

Stephen Council / Business Insider: Opsin, which helps companies show where AI agents are active, what data they access, and flag risky behavior, raised $21M led by SYN Ventures and Race Capital  —  AI agents are starting to outnumber humans in many high-tech workplaces, and they can be far trickier to monitor.

Discussion | Link

Two engineers shipped Stack Overflow for Agents in a few weeks with Codex, and its ChatGPT plugin helped drive up to 10× growth in agent creation.

Anant Kumar builds tools to compare loan valuations and analyze credit agreements, using Codex for code review and faster follow-ups.

The company generated over $30 million in revenue in two years and was acquired by MyFitnessPal for an undisclosed amount of money in March.

Reuters: Two US officials and a Western official say the order to limit Ukraine's access to SpaceX's Starshield came from inside the White House  —  The order to limit a satellite system used by Ukraine for battlefield communications and drone operations came from inside the White House …

Intero Biosystems beat out 199 other Startup Battlefield companies to take home the $100,000 grand prize at TechCrunch Disrupt 2026.

Mustafa Suleyman, the CEO of Microsoft AI. Stephen Brashear/Getty Images Microsoft AI recently announced its latest model MAI-Base-2. AI CEO Mustafa Suleyman compared the model to a similar offering from OpenAI. Microsoft trained its model more efficiently than OpenAI, he said in an internal meeting. Microsoft AI CEO Mustafa Suleyman told employees an in-house model was better than a similar offering from OpenAI, one of the software giant's biggest customers and partners. "We now have a stronger pre-trained base model than the model that currently powers GPT 5.6 Sol," Suleyman said during a town hall meeting with employees on Thursday. "And in fact, we used one-quarter of the compute that OpenAI used to train their model." Pre-training is the first stage of building an AI model, when it learns patterns from huge amounts of text, images, and other information. This takes massive amounts of expensive computing power, so Microsoft could save money if it's worked out a way to do this more efficiently. Microsoft announced the model, MAI-Base-2, earlier on Thursday. Suleyman said Microsoft privately tested this model against OpenAI models, including one nicknamed "Spud" that powers GPT-5.6. He added that Microsoft's model performed better on tests measuring how well AI can understand and predict patterns in information. The comparisons are based on technical information Microsoft has about OpenAI's models, Suleyman said. "Privately and confidentially, we've also compared it to the OpenAI models," Suleyman told employees, while suggesting that Microsoft has special access to information on how OpenAI models perform. It's notable that Suleyman would choose to highlight limitations of OpenAI models, even in private discussions. Microsoft has invested billions of dollars in OpenAI, and owns a stake in the startup that is likely worth more than $100 billion. Any criticism of OpenAI could potentially dent the value of this huge investment. The two companies also have a major partnership, with Microsoft offering OpenAI models to its Azure cloud customers. Microsoft has used OpenAI models for its own products such as Copilot, too. But the relationship between the two companies has frayed in the past year or so. Increasingly, Microsoft has tried to branch out from OpenAI, including by working on its own models. Suleyman also said Microsoft's new AI model used "12 times less computation than was used for DeepSeek, for example, on compute efficiency." Suleyman didn't specify which DeepSeek model, but a public blog post described the pre-training of MAI-Base-2 being a lot more efficient than a similar offering from the Chinese lab, called DeepSeek-V4-Pro-Base. Microsoft in March created a new combined team for its commercial and consumer Copilot products and reassigned Suleyman to focus on building AI models in the company's superintelligence unit. Microsoft declined to comment and OpenAI did not respond to a request for comment on Thursday. Have a tip? Contact this reporter via email at astewart@businessinsider.com or Signal at +1-425-344-8242. Use a personal email address and a nonwork device; here's our guide to sharing information securely. Read the original article on Business Insider

OpenAI announced on October 15, 2026 that its ChatGPT Chrome extension can now look across a user's open browser tabs, take control of those tabs to complete tasks, and organize and group tabs, extending its computer-use capability to an extension the Chrome Web Store lists at 7,000,000 users. In a series of posts on X from its official ChatGPT account, the company said it was bringing "the power of computer use to anyone who uses Chrome." In the announcement, OpenAI said that in Chat, the…

The world’s biggest equity rally is creating an unusual bout of FOMO among Taiwan’s professional stock pickers.

KDE celebrates 30 years with a new Plasma release, a farewell to X11, and a movie.

Attendees at OpenAI's DevDay conference on September 29, 2026. Minh Connors/Bloomberg via Getty Images OpenAI's Alexander Embiricos said the AI era gives a choice: work more, or work less. He's in charge of Dots, the company's new always-on work assistant. AI is often pitched as freeing up time. Embiricos responded to the idea that it might add more work. As the competitive AI race pushes tech workers into long hours, the head of OpenAI's new Dots product says AI gives people another option: work a whole lot less. Alexander Embiricos, who leads OpenAI's enterprise product and helms the new Dots agent tool, told a crowd at TechCrunch's Disrupt conference on Thursday that AI's capabilities have created something of a fork in the road for workers. You can grind to meet the moment. Or you can use AI to kick your feet up a little. "We basically have a lot more choice now," he said. Embiricos made the comments after a reporter asked about when AI systems will actually deliver more free time. The reporter shared an anecdote where an OpenAI employee used a ChatGPT mobile feature to work in their car and on a plane — not exactly liberated by the new tech. Embiricos said that the moment has already arrived where a worker can "outsource" tasks to AI and do less work. Then again, some people are so excited about AI's potential that they feel like it's an "interesting, ripe moment" to put in extra hours. "I think that's why a lot of people are motivated to work very hard right now. That includes myself," he said. "But I think actually there is this choice, and there's totally this other setup where you could delegate a ton more things." OpenAI is billing its Dots tool as a workplace assistant, always on and capable of proactively handing tasks and sorting to-dos. Embiricos uses his Dot often in Slack, he said, engaging on various threads. His perspective illustrates how early companies are in their journey to adopt AI tools. Most companies have not yet automated their work tasks, and many workers are still debating how much to use the tools. Early AI adopters, meanwhile, can sprint through some work that used to take days or weeks, especially in software engineering. Embiricos summed up the situation he feels workers are in: "You either choose: do you expand the ambition of what you can do, or do you do less? I do think we're there." Have a tip? Contact this reporter via email at scouncil@businessinsider.com, or over text, Signal, Telegram, or WhatsApp at 415-757-8198. Use a personal email address, a non-work WiFi network, and a non-work device; here's our guide to sharing information securely. Read the original article on Business Insider

The event closed in Toronto last week, with experts sharing critical perspectives on emerging risks, drawing record attendance and marking 22% growth.

The Close brings you the latest news and analysis leading up to the final minutes and seconds before and after the closing bell on Wall Street. Today's guests are Charles Schwab Chief Investment Strategist Liz Ann Sonders, IDC Senior Director, Data & Analytics Nabila Popal, Susquehanna Senior Equity Research Analyst Mehdi Hosseini, KPMG US Chair & CEO Tim Walsh, Envestnet CIO Dana D’Auria, Hirtle & Co Founder & Executive Chair Jon Hirtle, Gerber Kawasaki Wealth & Investment Management President & CEO Ross Gerber, Blue Sky Productions President John McElroy, & Porsche CEO Michael Leiters. (Source: Bloomberg)

Built from decades of threat intelligence research, the new resource explains scams from the victims' perspective and tracks emerging forms of cyber crime.