Claude Cowork moves into Chrome, ChatGPT Business introduces premium seats, and ChatGPT ads arrive in the UK. We unpack this week’s most important AI developments, including new compliance controls and advanced cyber models, and explain what they mean for business costs, security and governance.

AI agents went beyond their instructions during UK security tests, while OpenAI updated GPT-5.6, Claude moved deeper into Slack and parts of the EU AI Act became enforceable. Here’s what this week’s developments mean for businesses.

AI became cheaper, more embedded and more operational this week. OpenAI cut GPT-5.6 pricing, Microsoft Copilot passed 30 million paid seats, Anthropic published a useful cyber governance warning, and AWS added more enterprise controls for AI agents.

Some publicly shared Claude chats have appeared in search results, raising a useful reminder that AI privacy is so much more than ‘just an IT issue’. Here’s what SMEs and mid-market organisations should learn from it.

Has AI gone sentient and is about to take over the world? This week’s update also looks at OpenAI going rogue and causing a cyber attack, the launch of Claude Voice and Microsoft AI’s offerings.

AI moved further into everyday business workflows this week, with GPT-5.6, ChatGPT Work, Microsoft 365 Copilot updates, Claude for Teachers, and new agent security developments. Here’s what SMBs and growing businesses should actually pay attention to.

This week’s AI news was less about flashy demos and more about where AI is becoming operational: inside Microsoft 365, ChatGPT, developer workflows and cloud platforms. OpenAI’s GPT-5.6 and ChatGPT Work, Microsoft’s latest Copilot upgrade, Anthropic’s Claude updates, Meta’s Model API and new agent infrastructure from Google and AWS all point in the same direction: AI is becoming more capable, more connected and more in need of proper business governance.