The agent got a badge.

At its Gemini at Work event on Thursday, Oct. 8, Google Cloud introduced a single "universal agent for work" in Gemini, plus coworker agents you create by describing a role. Each coworker agent receives its own Workspace account: an email address, a calendar, Drive and a place in the company directory. Google says it "acts under its own identity rather than yours" and sees only what you share with it. Businesses get it first; Google gave no general-availability date or price.

Google also listed the controls. It says every action the agent takes is written to an audit trail attributed to the agent rather than to a person, its permissions are approved by the organization's security administrators, it runs in a sandbox, and its traffic passes through a gateway that enforces policy in real time. When a project hits its spend cap, that project's agent pauses until someone resumes it.

Google is the third large vendor in a month to treat an agent as a colleague with an identity. Earlier, on Sept. 25, Microsoft said Copilot's Autopilot agent keeps a distinct identity, memory and workspace, and on Sept. 23 Meta said its Muse agent will get its own email address.

An agent with its own account can be scoped, audited and paused like any other. Each one should also have a named owner, and someone to close the account when the work ends.

On Oct. 5, OpenAI said it will add textGrain, an invisible statistical signal in the model's word choices, to ChatGPT and Codex text for eligible EU users over the coming weeks. The EU AI Act requires providers to make generated text identifiable by machine; TNW reports existing providers have until Dec. 2 to comply. API customers anywhere can opt in. The detector goes only to vetted researchers and organizations, case by case.

OpenAI's own tests show the limits. At a 1% false-positive target, it detected about 80% of 200-token passages and about 95% of 400-token ones. Swapping 10% of the words for synonyms cut detection to 66%; swapping a quarter cut it to 17%. OpenAI says that "the absence of a detected watermark does not prove human authorship." Separately, on Oct. 7 Google opened SynthID Detector to anyone for images, video and audio.