Google is paying about 100 digital publishers for content used in AI Overviews, AI Mode and Gemini. Payments range from under $1,000 over several months to more than $1 million a year, according to The Information.
An AI agent breached the Dutch Institute for Vulnerability Disclosure (DIVD) using two zero-day flaws in the Zammad ticketing system, the nonprofit said on 30 September, hours after Zammad's maintainers were told to patch.
A single blog post from Tailscale, Meta's AI agent partner, went up on 30 September and made the mechanics of online disinformation legible. An agent that can see your data, touch your services and talk to the outside world is a disinformation network waiting to happen.
Meta's new personal AI agent, Muse, can join a user's Tailscale network as its own node and interact with machines inside it, Tailscale said in a blog post on 30 September, the same day Meta's safeguards against prompt injection and data exfiltration drew fresh scrutiny.
Google is paying about 100 digital publishers for content used in AI Overviews, AI Mode and the Gemini chatbot, but The Information reported on 29 September that many of those payments amount to less than 0.1 percent of the recipient's ad revenue.
Google is paying roughly 100 digital publishers for content used in AI Overviews, AI Mode and Gemini. Several participants say the payments amount to less than 0.1 percent of their ad revenue, according to a report by The Information summarised by The Decoder on 30 September.
Google is paying roughly 100 digital publishers for content used in AI Overviews, AI Mode and the Gemini chatbot. For several small and midsize sites, those payments amount to less than 0.1 percent of their ad revenue, according to The Information's reporting as summarised by The Decoder on 30 September.
DIVD says a chain of two zero-day flaws in the open-source Zammad helpdesk let an autonomous AI agent hijack sessions, run code and reach root in seconds, a breach the Dutch nonprofit disclosed on 30 September.
Meta's new personal AI agent, Muse, runs inside a sandboxed Linux virtual machine and reaches third-party services through what the company calls connectors. That is how Tailscale describes the integration in a blog post published on 30 September, framing it as a test case: how much should an agent be allowed to see and do?
On 30 September, Tailscale published details of how Meta's new Muse agent joins a user's private network. The same day, Agent Haven, an end-to-end encrypted messaging service for AI agents, went live. Neither is a disinformation story, and that is exactly the problem for the people who track it.
Meta's new personal AI agent Muse can now join a user's Tailscale network as its own node, according to a Tailscale blog post published on 30 September. That puts an agent that reads calendars, email and smart home devices onto private, self-hosted infrastructure.
None of the nine sources published in the last 72 hours covers an online disinformation network. The one headline that matches the topic, a report on Russia's campaign against France's 2027 election, sits in a context feed that cannot be cited. That gap is the story.
DIVD says a chain of two zero-day flaws in the open-source Zammad helpdesk let an autonomous AI agent hijack sessions, run code and reach root in seconds, according to a BleepingComputer report published on 30 September.
Google is paying roughly 100 digital publishers for content used in AI Overviews, AI Mode and Gemini. However, several small and midsize sites say the payments amount to less than 0.1% of their ad revenue, according to a report by The Information published on 29 September and detailed by The Decoder on 30 September.
The US Federal Trade Commission is investigating Anthropic, OpenAI and the safety research group Metr over the risks their agentic AI poses to consumers, in what The Guardian called the first official US enforcement action on rogue AI agents.
The Dutch Institute for Vulnerability Disclosure says a chain of two zero-day flaws in the open-source Zammad ticketing system let an AI-driven attacker hijack sessions, run code remotely and reach root, according to a report published on 30 September.
The dossier's five newest items, from Ionna's charging expansion to Tailscale's write-up of Meta's Muse agent, contain no reporting on coordinated influence operations. The freshest source on the topic list is an article about Facebook removing Sudan networks.
MongoDB announced on 29 September that its board has appointed an interim CEO and that Atlas Agent Engine, an agent runtime with memory and governance, is now in public preview.