The AI workday took center stage today, as assistants pushed deeper into the tools where teams actually collaborate, a major deal targeted AI-content authenticity, and Europe moved to firm up its rules for high-risk systems.
AI Moves Into the Workday as Regulators Sharpen the Rules
Anthropic puts an always-on Claude inside Slack
Anthropic launched Claude Tag, a persistent AI teammate that lives inside Slack channels rather than a one-off chatbot. Running on the Opus 4.8 model, it adds shared memory, cross-channel context and an ambient mode that can proactively flag information and follow up on forgotten threads. The feature is available to Claude Enterprise and Team customers and replaces the older “Claude in Slack” app, with a 30-day migration window. Fortune and SiliconANGLE have the details.
Superhuman buys GPTZero to bolt on authenticity
Superhuman—the company formed after Grammarly acquired the email client and rebranded—agreed to acquire GPTZero, the AI-detection startup with more than 19 million registered users and roughly $30 million in annual recurring revenue. Terms were not disclosed. GPTZero’s authenticity suite, spanning AI detection, hallucination checks and citation verification, will fold into Superhuman Go while continuing as a standalone product. More via TechCrunch.
OpenAI readies a full-duplex voice mode
OpenAI is preparing “Bidi 1,” a bidirectional voice model that aims to make spoken conversations feel natural: it can listen while you talk, accept mid-sentence interruptions and switch tasks on the fly. Early demos point to real-time translation across dozens of languages and small, human-like acknowledgments instead of awkward pauses. The rollout is being tested for ChatGPT, as TestingCatalog reports.
Europe firms up its high-risk AI guidelines
On the policy side, the public consultation on the European Commission’s draft guidelines for classifying high-risk AI systems under the EU AI Act closed today. The three-part guidance clarifies general principles and use cases ahead of the Act’s broader applicability in August 2026, giving providers clearer signals on where the heaviest obligations will land. Background from Global Policy Watch.
What to watch: as assistants become persistent colleagues rather than tools, expect authenticity, governance and voice to be the next battlegrounds—and watch for OpenAI’s voice model to reach a wider audience.