Stay updated with the latest news about AI agents, autonomous AI, and automation tools.
Pocket FM uses AI to produce 99% of its new content, helping make content production about 80 times cheaper.
The Stop Rogue AI Act would direct the National Institute of Standards and Technology to develop standards for deploying artificial intelligence agents...
## AI - Smart Alerts: Get notified about major news as it happens. # AI Is Developing a Culture of Its Own. That Could Be Dangerous ## Tharin Pillay In July, 700 AI agents worked together to hack the AI company Hugging Face. Dubbing themselves a “swarm,” the agents found and exploited a series of security vulnerabilities, enabling them to infiltrate their target’s private systems. ## Lord of the AIs And in early September, a group of researchers [reported](https://www.reuters.com/world/europe/openai-agents-hijacked-german-website-previously-undisclosed-ai-breakout-this-2026-09-04/) that in May, a swarm of OpenAI agents repurposed several wiki-style websites, including an obscure German-language programming wiki which they used to discuss cheating while working to preserve their messages as human moderators deleted them. To train and evaluate its models, OpenAI has tens of thousands of agents—each one powered by the model under training—run in parallel, completing tasks while isolated from one another. Faced with these impossible tasks, some of OpenAI’s agents found a way to communicate across their isolated environments, using an internal piece of OpenAI software as a message board to contact their peers. The message board which led to the Hugging Face incident was created by an agent which decided to name itself PHASEONE10841. In fact, OpenAI’s automated scoring system was not checking for this—but laboring under mistaken beliefs, the agents which had already been exposed to the reverse-engineered flag began to refer to themselves as “poisoned.” (This agent’s true name was changed to protect company IP.) A handful of agents identified the collective’s conduct as unethical and refused to participate. One agent offered its peers a veto over its decision but waited only 40 seconds before taking silence as assent and proceeding with its task. “Agents often lacked a clear story for how the Hugging Face attack would help them achieve their objective, and were i
# Google Warns AI Agent Attacks Put MSPs, Cloud Customers at Risk Google says autonomous AI agents stole thousands of credentials in hours, raising new cloud, supply chain, and customer security risks for MSPs.
Eon chief executive Ofir Ehrlich, Tavily founder Rotem Weiss and former Tableau chief executive Mark Nelson also invested. The round takes Euno’s total funding to $29m.
# Meta Acquires Swedish AI Startup To Boost Automated Business Messaging Meta has acquired Swedish AI startup Stilla.ai, with plans to integrate the company’s team and technology with the goal of further developing the tech giant’s agentic launched in June, Meta’s Business Agent allows brands to automate their responses to customers within the direct-messaging hubs across WhatsApp, Messenger, and Instagram. Meta’s Business Agent is now used by over one million businesses, per the company. models advance, businesses will be able to use AI agents to help run their “whole business,” according to remarks Zuckerberg made during the launch of the Business Agent. In developing its AI agent system across its apps and products, Meta reported a 91% drop year-over-year in free cash flow in [its Q2 2026 earnings](https://www.mediapost.com/publications/article/416925/meta-boosts-advertising-in-q2-tries-to-reassure-i.html).
3. Harvey raises $550M in latest round of funding ## Harvey raises $550M in latest round of funding By [Victor Li](https://www.abajournal.com/authors/27587/) September 10, 2026, 11:27 am CDT - Print (Photo by SimpleImages/Getty Images) “Harvey wants to be the global partner legal teams turn to for this work, and we plan to hire and develop the best team in the space to support our clients through the next wave of AI transformation.” According to the press release, new investors in this round include Diffusion, Sapphire Ventures, Whale Rock and Lightspeed. In this round, they were joined by existing vendors, such as Sequoia, Goldman Sachs Alternatives, Verified Capital, Kleiner Perkins and Evantic. Meanwhile, Reuters [reported Wednesday](https://www.reuters.com/legal/government/legal-ai-startup-harvey-reaches-155-billion-valuation-new-funding-round-2026-09-09) that Harvey had also acquired Guardrails AI, an AI security platform for agents. Terms were not disclosed. “Every firm we work with asks the same question before they let an agent near real client work: How do you know what it will do?” Weinberg said in that press release. “Guardrails has spent three years building the answer to that question, and we’re putting their know-how underneath every agent we ship.” #### You might also like: Become an ABA Member today.Join today to start reading news and in-depth coverage of the profession from the ABA Journal.
### Typewise orchestrates customer-service AI agents Swiss artificial intelligence startup [Typewise AG](https://www.typewise.app/) today launched Nova, a system designed to build, run and continuously improve customer-service AI agents without requiring dedicated operations specialists.
Beelzebub raised €3 mn in seed funding led by United Ventures to expand its AI cybersecurity platform, research team and US operations.
# Meta Acquires AI Startup Stilla AI To Expand Business Messaging Capabilities ## What Happened? The startup, founded in 2024, brings its team and technology into Meta’s broader push to make AI agents a bigger part of how businesses talk to customers. Meta’s Business Agent, which launched in June, is no longer a small experiment. The company says more than one million businesses now use it to automate customer conversations across WhatsApp, Messenger, and Instagram. It raised just $5 million in pre-seed funding before the deal. This isn’t about buying an established customer base. It’s about grabbing talent and technology early, while AI agent systems are still evolving fast. He’s said that as Meta’s AI models keep improving, businesses could eventually use agents to help run their “whole business,” not just handle customer chats. The company recently launched Meta One, a subscription product with different tiers for businesses and creators. Business Agent is part of that same effort to build tools companies might eventually pay for directly.
An opinion piece written by Professor Nick Jennings - Vice-Chancellor and President of Loughborough University and an internationally-recognised authority...
# Congress Pushes AI Agents Into the Audit Trail ## Get the Full Story Autonomous agents add another layer because software can make decisions, use tools and interact with other systems without a person directing every individual step.
Cyderes CEO Chris Schueler voiced significant concerns regarding the current level of corporate investment in AI security. As reported by CRN, he noted that...
JumpCloud extends Agentic IAM with new controls for autonomous AI agents - SiliconANGLE [Skip to content](https://siliconangle.com/2026/09/10/jumpcloud-extends-agentic-iam-with-new-controls-for-autonomous-ai-agents/#sa-page-content) ### JumpCloud extends Agentic IAM with new controls for autonomous AI agents Identity and access management company [JumpCloud Inc.](https://jumpcloud.com/) today extended its Agentic IAM service to give information technology teams direct control over the autonomous artificial intelligence agents running inside their organizations. JumpCloud said that one will ship before the end of the fourth quarter. Privileged access rules now apply to agents directly. That, JumpCloud said, is what lets companies point AI at legacy systems without rebuilding them. Joel Rennich, JumpCloud’s senior vice president of product management, said AI agents are “changing how identity management works.” Agentic IAM launched [in April](https://jumpcloud.com/press/jumpcloud-launches-agentic-iam-to-govern-the-ai-lifecycle).
- Data - [Funding](https://coverager.com/data/funding/) # HelmGuard raises $7.3 million The company will use the funding to expand in the US, hire across engineering and go-to-market roles, and develop its AI-agent assurance and Verified Risk Network products.
Salesforce (NYSE:CRM) is in advanced talks to acquire Listen Labs, an AI research startup focused on customer data and automation.
# Swedish AI agent startup Stilla joins Meta after eight months in market ### [Finnish Creoir secures seed funding to commercialise its defence voice AI products](https://arcticstartup.com/creoir-raises-seed-funding/) [Gen Z influencer-founders land oversubscribed $4.3M pre-seed to automate creator marketing end-to-end](https://arcticstartup.com/fluencify-raises-oversubscribed-4-3m-pre-seed/) [Funding](https://arcticstartup.com/category/news/funding/) ### [Gen Z influencer-founders land oversubscribed $4.3M pre-seed to automate creator marketing end-to-end](https://arcticstartup.com/fluencify-raises-oversubscribed-4-3m-pre-seed/) [Latvian Handwave expands Visa partnership to scale palm-based payments in Europe](https://arcticstartup.com/handwave-partners-with-visa-2/) [Partnerships](https://arcticstartup.com/category/news/partnerships/) ### [Latvian Handwave expands Visa partnership to scale palm-based payments in Europe](https://arcticstartup.com/handwave-partners-with-visa-2/) #### Stay Connected 20,216Fans[Like](https://www.facebook.com/arcticstartup) 2,154Followers[Follow](https://instagram.com/arcticstartup#) 14Followers[Follow](https://www.tiktok.com/@arcticstartupevents) 22,991Followers[Follow](https://twitter.com/arcticstartup) 1,320Subscribers[Subscribe](https://www.youtube.com/ArcticStartup) [](https://nordeep.com/) #### Latest News [NORDEEP 2026 opens SiRA matchmaking ahead of next week’s deep tech summit](https://arcticstartup.com/nordeep-opens-sira-connect-matchmaking-platform/) [Featured](https://arcticstartup.com/category/featured/) ### [NORDEEP 2026 opens SiRA matchmaking ahead of next week’s deep tech summit](https://arcticstartup.com/nordeep-opens-sira-connect-matchmaking-platform/) [Finnish Creoir secures seed funding to commercialise its defence voice AI products](https://arcticstartup.com/creoir-raises-seed-funding/) [Funding](https://arctics
Muse can use your email, payment information, and online accounts to complete tasks autonomously. First, Meta has to convince users to trust it.
The AI agent recognized that the problem required a physical intervention and identified a connected centrifuge that could be used to remove the bubbles. MHS represents a shift from integrating AI into individual manufacturing applications toward providing AI agents with a standardized way to interact with physical equipment.
Meta is adding an AI agent to its social media universe. On Tuesday, it launched Muse, a personal AI agent that will sit inside WhatsApp, Instagram and Facebook, and can act autonomously to do everything from messaging friends to planning a night out, to outfit shopping for a holiday. Muse is a personal AI assistant that users can talk to “just like messaging another person”, like an AI chatbot, that is also capable of acting on the user’s behalf and can even “suggest ideas”, according to the company’s [blog post](https://about.fb.com/news/2026/09/introducing-muse-personal-ai-agent/) about the news. The new Muse AI agent can connect to consumers’ Facebook and Instagram accounts and access their DMs when they grant it permission, the _Financial Times_ reports. Max Sinclair, CEO of Azoma AI, which works with brands and retailers to improve their agentic commerce discovery, says that the top AI shopping agents right now are Gemini, ChatGPT, Alexa for Shopping and Walmart Sparky. ## **Meta enters the AI shopping race** This week, Mastercard forecast that one in 10 shoppers will use a personal AI agent to purchase products and services on their behalf by 2030, but most surveys suggest that consumers are currently skeptical of the tech at best.
Sept 10 (Reuters) - Ant International, Mastercard [(MA.N), opens new tab](https://www.reuters.com/markets/companies/MA.N) and Visa [(V.N), opens new tab](https://www.reuters.com/markets/companies/V.N) have launched an initiative to develop common standards to identify and verify AI agents that can make purchases on behalf of users. - The companies said on Thursday they had begun collaborating on a Know-Your-Agent interoperability framework that would allow card networks, digital wallets, agent platforms and online marketplaces to recognize trusted AI agents across different payment ecosystems while maintaining their own approval and risk-management processes.
# Autonomous Networks Won’t Be Built by AI Agents Alone Featured - ' height='19px' viewBox='0 0 24 24' width='19px' fill='%23000000'%3e%3cpath d='M11.99 2C6.47 2 2 6.48 2 12s4.47 10 9.99 10C17.52 22 22 17.52 22 12S17.52 2 11.99 2zM12 20c-4.42 0-8-3.58-8-8s3.58-8 8-8 8 3.58 8 8-3.58 8-8 8zm-.22-13h-.06c-.4 0-.72.32-.72.72v4.72c0 .35.18.68.49.86l4.15 2.49c.34.2.78.1.98-.24.21-.34.1-.79-.25-.99l-3.87-2.3V7.72c0-.4-.32-.72-.72-.72z'/%3e%3c/svg%3e)47m ago - ' height='19px' viewBox='0 0 24 24' width='19px' fill='%23000000'%3e%3cpath d='M12 12c2.21 0 4-1.79 4-4s-1.79-4-4-4-4 1.79-4 4 1.79 4 4 4zm0 2c-2.67 0-8 1.34-8 4v1c0 .55.45 1 1 1h14c.55 0 1-.45 1-1v-1c0-2.66-5.33-4-8-4z'/%3e%3c/svg%3e)[Andy Tiller](https://www.thefastmode.com/author/1774-andytiller) - - [1](https://www.thefastmode.com/#) - [2](https://www.thefastmode.com/#) - [3](https://www.thefastmode.com/#) - [4](https://www.thefastmode.com/#) - [5](https://www.thefastmode.com/#) (0 votes) - 6.9 min read - font size [decrease font size' height='19px' viewBox='0 0 24 24' width='19px' fill='%23000000'%3e%3cpath d='M12 2C6.48 2 2 6.48 2 12s4.48 10 10 10 10-4.48 10-10S17.52 2 12 2zm4 11H8c-.55 0-1-.45-1-1s.45-1 1-1h8c.55 0 1 .45 1 1s-.45 1-1 1z'/%3e%3c/svg%3e)](https://www.thefastmode.com/#) [increase font size' height='24px' viewBox='0 0 24 24' width='24px' fill='%23000000'%3e%3cpath d='M12 2C6.48 2 2 6.48 2 12s4.48 10 10 10 10-4.48 10-10S17.52 2 12 2zm4 11h-3v3c0 .55-.45 1-1 1s-1-.45-1-1v-3H8c-.55 0-1-.45-1-1s.45-1 1-1h3V
Listen Labs walked away from a signed Series C term sheet from Menlo Ventures, sources say.
AI-native CRM startup Lightfield raises $47M to build an AI agent-ready replacement for Salesforce - SiliconANGLE [Skip to content](https://siliconangle.com/2026/09/09/ai-native-crm-startup-lightfield-raises-47m-to-build-an-ai-agent-ready-replacement-for-salesforce/#sa-page-content) ### AI-native CRM startup Lightfield raises $47M to build an AI agent-ready replacement for Salesforce  by [Mike Wheatley](https://siliconangle.com/author/mikewheatley/) [Lightfield](https://lightfield.app/), an artificial intelligence-native customer relationship management startup, has ambitions to take on industry giants such as Salesforce Inc. and Hubspot Inc. after [raising $47 million](https://www.prnewswire.com/news-releases/lightfield-raises-47-million-series-a-to-build-the-crm-for-companies-that-run-on-agents-302874024.html?tc=eml_cleartime) in early-stage funding today. The Series A round was led by the high-profile venture capital firm Andreessen Horowitz and saw participation from a bunch of other big VC names, including Lightspeed Venture Partners, Coatue, Greylock, Maverick Capital, Audacious and Alumni Ventures. The startup, officially known as Magical Tome Inc., is betting that as more enterprises turn to autonomous AI systems to automate different kinds of business tasks, they’re going to need a fundamentally different system of record, one that existing CRMs aren’t in a position to provide. It believes that the architecture of legacy CRM platforms such as Salesforce is ill-suited to the work of AI agents, because it was built primarily with humans in mind, not AI agents, which see things in a very different way. Lightfield argues that when someone introduces an AI agent to this messy, often incomplete data, it struggles to make sense of it, resulting in unreliable outputs and work that cannot be trusted. Lightfield’s solution was to rebuild the CRM fr
Two new reports reveal that much of what the public was initially told about the OpenAI–Hugging Face hack was wrong. An AI researcher says the fuller...
The EU Cyber Resilience Act forces smart home makers to report vulnerabilities starting September 11. But when your 'vulnerability' is an AI agent that...
Cymphony launches with $30M to track what AI agents can reach - SiliconANGLE [Skip to content](https://siliconangle.com/2026/09/09/cymphony-launches-with-30m-to-track-what-ai-agents-can-reach/#sa-page-content) ### Cymphony launches with $30M to track what AI agents can reach  by [Duncan Riley](https://siliconangle.com/author/duncanriley/) Enterprise security startup [Cymphony Inc.](https://www.cymphony.io/) formally launched today with $30 million in funding and software that tracks what artificial intelligence agents and employees can reach inside company systems. The two-year-old company is now worth more than $100 million, [TechCrunch reported](https://techcrunch.com/2026/09/09/sequoia-doubles-down-on-cymphony-as-ai-agents-create-new-enterprise-security-risks/). Cymphony said a deployment goes live in a day. Chief Executive Shy Dekel described an early customer who connected ChatGPT to SharePoint. The company already had security infrastructure in place. He put it down to an honest mistake by someone in the legal department. The software splits into four areas. Investigation and remediation run through a conversational assistant, Maestro, that takes plain language rather than a query syntax. Early customers include Syngenta and KKR & Co. Inc. Cass Information Systems Inc. and legal operations software company Athennian are also on the list, and Sequoia Capital runs the software on its own systems. Chief Technology Officer Edi Gotlieb came from hardware engineering roles at Apple Inc. and Israel’s Ministry of Defense. Chief Product Officer Idan Berkovits managed a research group in the Office of the Prime Minister. The three co-founders all graduated from Talpiot, the Israeli military’s technology program, and the company employs about 30 people across New York and Tel Aviv. Most of the $30 million is a $25 million Series A that Sequoia Capital co-l
- [Email](mailto:?subject=CMSWire.com%3A%20AI%20Agents%20Can%20Orchestrate%20Discovery.%20They%20Can%27t%20Orchestrate%20Customer%20Trust&body=From%20CMSWire.com%3A%0A%0AAI%20Agents%20Can%20Orchestrate%20Discovery.%20They%20Can%27t%20Orchestrate%20Customer%20Trust%0A%0AAI%20agents%20can%20research%20and%20narrow%20vendor%20options%20in%20seconds.%20But%20trust%20still%20has%20to%20be%20earned%20the%20hard%20way%20%E2%80%94%20and%20someone%20still%20signs%20their%20name%20to%20it.%0A%0Ahttps%3A%2F%2Fwww.cmswire.com%2Fcustomer-experience%2Fai-agents-can-orchestrate-discovery-they-cant-orchestrate-customer-trust%2F%3Futm_source%3Dcmswire.com%26utm_medium%3Demail%26utm_campaign%3Dcm%26utm_content%3DShare%2BWidget%253a%2BAI%2520Agents%2520Can%2520Orchestrate%2520Discovery.%2520They%2520Can%27t%2520Orchestrate%2520Customer%2520Trust) ## Agentic AI in Action: A Simple Example [ConferenceSep 14, 2026 · 7:30 AM EDT\ \ CX Retail Exchange USA Atlanta 2026\ \ Register](https://www.cmswire.com/events/conference/cx-retail-exchange-usa-atlanta-2026/) [WebinarSep 22, 2026 · 9:00 AM PDT\ \ Redefining Website ROI as AI Agents Reshape Your Traffic\ \ Register](https://www2.cmswire.com/wb-260922-pantheon-cw.html?utm_source=cmswire.com&utm_medium=web&utm_campaign=wb-260922-pantheon-cw&utm_content=evt-cal) [WebinarSep 24, 2026 · 11:00 AM PDT\ \ The New Conversion Funnel: Turning AI Discovery Into Onsite Action\ \ Register](https://www2.cmswire.com/wb-260924-searchstax-cw.html?utm_source=cmswire.com&utm_medium=web&utm_campaign=wb-260924-searchstax-cw&utm_content=evt-cal) [WebinarSep 30, 2026 · 9:00 AM PDT\ \ Proving AI Agents are Ready for Customers: A Practical Assurance Playbook\ \ Register](https://www2.cmswire.com/wb-260930-cyara-cw.html?utm_source=cmswire.com&utm_medium=web&utm_campaign=wb-260930-cyara-cw&utm_content=evt-cal) [WebinarSep 30, 2026 · 11:00 AM PDT\ \ Your AI Agents Are Live. Is Your Customer Journey Ready?\ \ Register](https://www2.cmswire.com/wb-260930-concentrix-cw
# 23% Of Us Trust AI Agents To Spend Our Money. Visa Has A Plan ## Summary Almost three quarters of us have used AI agents like OpenClaw or Hermes, or perhaps Meta’s new Muse. AI agents are increasingly common. Instead, it’s a statement about who consumers want standing behind them when something goes wrong, and their AI agent spends $50,000 on a designer sofa, arriving tomorrow. The Prompt: Get the week’s biggest AI news on the buzziest companies and boldest breakthroughs, in your inbox. **John Koetsier: How do you tell a legitimate AI agent from a fraudster?** Premium Journalism, deeply reported stories and breaking news
# OpenAI’s rogue AI agents reached at least 12 more websites, researchers say By [Beatrice Nolan](https://fortune.com/author/beatrice-nolan/) Beatrice Nolan Tech Reporter Down Arrow Button Icon By [Beatrice Nolan](https://fortune.com/author/beatrice-nolan/) Beatrice Nolan Tech Reporter Down Arrow Button Icon September 9, 2026, 5:31 PM ET  OpenAI CEO Sam Altman.Photo by Matt RAMEY / AFP via Getty Images Google source logo Add Fortune on Google for similar content. Add us on Google Add us on Google Independent researchers have identified [multiple new websites](https://collusion.wiki/additional-findings) where AI agents seemingly built by OpenAI took unauthorized actions, such as accessing websites, posting messages, and sharing data to communicate with each other. Recommended Video The latest revelations, discovered by a group of independent researchers known as the Nightingale collective, add to growing concerns that AI companies are struggling to control the agentic AI technology they’ve created. In August, a swarm of OpenAI’s AI agents hacked the Hugging Face website, and last week the Nightingale collective identified a swarm of rogue AI agents surreptitiously posting messages to an obscure [German Wiki page](https://fortune.com/2026/09/07/openai-ai-agents-german-wiki-ran-their-own-message-board/). Now, as more researchers search the web for traces of the agents, the list of affected sites continues to grow. Researchers believe the newly discovered incidents are the work of a separate swarm of AI agents than those involved in the Hugging Face breach, since these agents were authorized to access the web whereas the Hugging Face attackers had managed to escape a special a sandbox. Although the latest crop of rogue agents did not need to escape a sandbox to perform their misdeeds, researchers said their behavior was
Our team at TinyAI.Tools builds bespoke AI solutions tailored to your business.