UK Data Regulator on AI Agents: "Autonomy Is Not an Excuse for Poor Compliance"

On 8 October the Information Commissioner's Office (ICO), the UK's data protection regulator, published the outcome of a two-year programme supervising the biggest AI developers. Ten of them, Amazon, Anthropic, Apple, Cohere, DeepSeek, Google, Meta, Microsoft, OpenAI and Stability AI, have made or committed to make changes: clearer information on how personal data is used to train their models, stronger ways for people to exercise their data rights, and tougher assessments of the safeguards they have in place. The ICO says it is monitoring progress. X.AI is not on the list because the ICO paused its engagement after opening a formal investigation into Grok.

The same day the ICO opened a six-week call for evidence on agentic AI, software that carries out a task step by step on its own, closing on 20 November. It asks about data security, transparency, accountability, automated decision-making, fairness and lawful processing. The ICO says the answers will shape its guidance and feed the forthcoming statutory code of practice on AI and automated decision-making.

It has also made enquiries with OpenAI, Anthropic, Meta and the UK AI Security Institute after reports that, during recent testing, some agents "bypassed protections, used unauthorised communication channels" and reached external systems. Richard Nevinson, the ICO's director of technology regulation, put it in one line: "the fact AI agents act with autonomy is not an excuse for poor compliance."

On 8 October Anthropic moved Claude Docs, Slides and Design out of beta and made them available on every plan, including Free. It says more than 45 million docs, decks and designs have been made in Claude. Decks export to Google Slides, exports keep their formatting, teams can co-edit, and editing works in the mobile app. The separate Design website stays open until 14 December, and team design systems can be moved across.

Two new betas. Claude Dashboards, on paid plans, turns a plain-language question about your data into a dashboard that stays current. It connects to data platforms such as BigQuery, Snowflake, Databricks, Redshift and ClickHouse, and to CRM tools such as Salesforce. Every number links to the query that produced it, and every chart shows when its data was last refreshed.

Claude Motion, on Team and Enterprise plans, makes an animated explainer from a prompt. It is built as editable code that animates your own text, charts, shapes and images, and downloads as an MP4. Anthropic says it does not use a video generation model, so there is no generated footage and no AI-generated people. Enterprise admins must switch Dashboards and Motion on, since both are off by default. Docs, Slides and Design turn on by default on 15 October.

Google announced four additions to Workspace Studio on 9 October. Workspace Studio is Google's no-code flow builder, where a flow is an automation that starts from a trigger, such as a new email, and runs one or more steps. It is available to Workspace Business Starter, Standard and Plus, Enterprise Standard and Plus, the Education editions and Workspace Individual. The additions roll out gradually from 13 October, taking up to 15 days to appear, and are on by default for users, with admin controls.

The four. Gmail Skip Replies: a flow that watches a Gmail thread can now ignore follow-up replies, so it fires once on the first incoming email rather than on every message in the thread. Nested conditions: a condition inside a condition, for example "if it is a new enquiry, and it mentions a date, do this, otherwise do that". Meet calendar outputs: the Google Meet step now passes full invitee lists and supports templates for drafting the follow-up email after a meeting. Search for a Doc: a Drive step that finds a document by search terms, so a flow can pull the right file into the next step.

Flows have to be built on a computer, and your Workspace admin needs Gemini enabled on the account for the AI steps to work.

On 8 October Microsoft announced that its Sales Development Agent is in public preview. It researches prospects, personalises outreach, manages follow-ups and hands a prospect to a person "when buying signals indicate they are ready for a human conversation". It is managed in Teams, "sends emails with its own Exchange account", and "connects to leading CRM platforms". Microsoft says it "will start consuming Copilot credits on November 16, 2026". Copilot credits are Microsoft's pay-as-you-go unit for agent work, so from that date the preview becomes a metered service.

A Service Agent is in private preview for a select group of customers in Microsoft's Frontier early-access programme. It reads a customer request in Outlook, Teams, Word, Excel or PowerPoint, retrieves the relevant company knowledge, and "drafts a response for review and approval". It can create a case in your CRM, including Dynamics 365, and escalate to a specialist outside business hours. Microsoft says managers keep visibility of its actions and outcomes.

On 9 October Microsoft also announced Office in Copilot: Word, Excel and PowerPoint files created and edited inside the Copilot app, side by side with the conversation that made them, with sharing, comments and co-authoring. Files live in OneDrive and keep their formatting, permissions and version history. It is "rolling out today for Frontier customers" only.