Anthropic’s Claude Now Crashes Google Meet Calls to Do Your Job

Anthropic has rolled out a beta test dubbed Claude Meetings, turning its chatbot into an autonomous participant inside Google Meet rooms. Instead of just lurking silently like half the marketing team, the model actively listens, scribbles transcripts, and handles real-time assignments thrown its way.

Attendees can prompt the AI via vocal commands or directly inside the meeting chat, though it politely defaults to text responses. When put to work during live brainstorming, it can instantly spin up interactive whiteboards, draft documentation, pull last year’s expense sheets, and compare them against current budgets while humans are still busy arguing about calendar invites.

To fetch sensitive records, the bot hooks directly into connected workspace integrations, requiring explicit authorization from specific users. Because it is not an all-seeing oracle quite yet, vocal requests trigger an identity verification check where it clarifies exactly who is issuing the command before touching private company data.

Once the camera goes dark, Claude compiles a dedicated dashboard featuring full verbatim transcripts, summarized key points, action items, and every artifact created during the session. Links are blasted out to everyone's inbox, with future rollouts slated to bridge data straight into third-party tools like Plaud, Gong, and Granola. The beta is currently gated behind Team and Enterprise tiers across desktop and web clients, tallying integration queries against standard user rate limits.

The corporate illusion of contributing during endless calls is swiftly collapsing. Having an unflinchingly attentive silicon coworker cross-referencing past budgets and taking verbatim notes will make slacking off during weekly retrospectives a lost art form.

Source: Anthropic