Today's Headlines

  • Google Cloud turns Gemini into an agent for work, including coworker agents inside Workspace
  • Anthropic rewrites its Usage Policy with a new section on deceptive campaigns, effective November 12
  • Manus's parent raises more than $500 million, its first round since the Meta deal was unwound

Today's three stories are about what to hand to AI and where to draw the line. Google Cloud presented Gemini as an agent you give an objective to and that comes back with finished work.

Anthropic rewrote the uses it rules out for Claude, from fake-account influence campaigns to tracking people without consent. In China, Manus closed its first funding round since authorities ordered its sale to Meta unwound.

Today's Top Three

Google Cloud turns Gemini into an agent for work, including coworker agents inside Workspace

Google Cloud announced the Gemini agent, which it calls a single, universal agent for work, at its Gemini at Work 2026 event on October 8.

From one prompt box, it answers questions, handles knowledge work, creates images and media, and writes and runs code. Google Cloud CEO Thomas Kurian said users give it "objectives, not instructions." Gemini plans the work, picks the skills and tools it needs, connects to company systems, and returns finished output.

It also works directly inside Gmail, Drive, Docs, Sheets, Calendar and other Workspace apps. Teams can set one up as a coworker agent with its own Workspace account, including an email address and calendar, and colleagues assign it work by mentioning it in a chat. Every action it takes is written to an audit trail attributed to the agent itself.

Gemini picks a model for each job, orchestrating across Google's Gemini models and Anthropic's Claude models today. Companies can set a hard spending cap per project, and the agent pauses when the cap is reached. Versions tailored to financial services and legal teams are in preview.

Google says nearly 90% of the Fortune 100 use Gemini Enterprise. SOMPO, the Japanese insurer, has built more than 10,000 custom agents across its 34,000 employees. According to The Verge, the Gemini agent is currently available to enterprise customers in private preview.

Anthropic rewrites its Usage Policy with a new section on deceptive campaigns

Anthropic published a new version of its Usage Policy for Claude on October 8. It takes effect on November 12.

A new section, Do Not Engage in Deceptive Campaigns or Artificial Activity, covers deceptive activity of any kind, political or commercial: hiding who is behind a message, amplifying content through fake accounts or posts, and building the tools that run influence operations. It consolidates rules that were previously scattered across the elections, fraud, privacy and disinformation sections. The elections section now focuses on deceiving voters and disrupting elections, and has been renamed Do Not Undermine Democratic Processes.

The surveillance and criminal justice section was rewritten. Tracking people without their consent is prohibited, whether in real time or through analysis of previously collected data, and Claude cannot be used to decide or recommend who to investigate, arrest or charge. The section also lists what remains permitted, including consented tracking such as fraud monitoring, content moderation, journalism and legal research.

The weapons section now spells out that the ban covers the software and components that make weapons work, as well as arming drones and other autonomous vehicles. For high-risk uses that affect health, legal rights, finances or livelihoods, the requirements for a qualified human in the loop and for telling people that AI was used remain in place, and the section now lists which kinds of recommendations they cover. When Claude drives hardware that takes autonomous physical actions, an operator must be able to observe and stop the equipment, and it must hold a safe state if Claude disconnects.

The policy also prohibits sustained and needless abusive or cruel behavior toward Anthropic's models, which Anthropic says applies only in extreme cases and covers neither ordinary frustration nor dark creative themes nor model testing. The Supported Regions Policy now states that Claude may not be used by people located in unsupported regions, by entities incorporated or headquartered there, or by entities majority-owned or controlled from there.

Manus's parent raises more than $500 million, its first round since the Meta deal was unwound

Butterfly Effect, the parent of Chinese AI agent company Manus, said on WeChat that it has closed a funding round of more than $500 million.

Boyu Capital and IDG Capital led the round, with existing shareholders Tencent, HongShan (formerly Sequoia China) and ZhenFund also taking part. The company did not disclose its valuation. According to TechCrunch, Manus was reported last month to be in talks to raise $500 million at a $4 billion valuation.

TechCrunch reports that Manus agreed in December 2025 to be acquired by Meta for $2 billion. Chinese authorities ordered the deal unwound in April 2026, amid growing concern in China about losing AI talent to the West, and Manus resumed independent operations in August. This is its first funding round since then.

Manus says it is continuing to hire in China and abroad. TechCrunch adds that the company is reported to be considering a listing in Hong Kong.

Other Developments

Policy

  • OpenAI said it banned ChatGPT accounts tied to two covert influence operations, one from Russia and one from Iran. The Russian operation appears to have had unwitting people in Latin America run a "think tank" for it, and OpenAI rated it Category 5 on a six-point breakout scale, the first Category 5 operation it has disrupted since it began reporting. The Iranian operation pitched articles to small and medium online outlets under seven fake journalist bylines. Disrupting AI-enabled "false front" operations (OpenAI)
  • USA Today Co. and several of its local newspapers sued OpenAI, alleging it copied "hundreds of thousands" of articles to train its models, The Verge reports. The publisher is seeking more than $250 million in damages. USA Today becomes the latest publisher to sue OpenAI (The Verge)
  • Anthropic launched the Anthropic Cyber Mission to support defenders of critical systems. Its Critical Infrastructure Defense Program has 11 founding partners, including Hitachi, CrowdStrike, Palo Alto Networks and Accenture, and its new OSS Scanner gives open-source projects regular security scans for free. Introducing the Anthropic Cyber Mission (Anthropic)

Research

  • Anthropic committed $150 million over three years to the Genesis Mission, a US federal initiative to speed up scientific discovery with AI. The funding will make Claude available to more than 15 agencies in the mission, including NASA, the National Institutes of Health and the National Science Foundation. Building on our commitment to American scientific discovery (Anthropic)

Business

Source: Selected by the editors from the AI news inbox (collected October 9, 2026; 93 items, 27 primary and 66 secondary).