The regulator published its findings on 8 October and opened a call for evidence on AI agents, which stays open until 20 November. Here is what it means for a small firm.

By Jason Sibley
On 8 October 2026 the Information Commissioner's Office (ICO) published a report on ten of the biggest foundation model developers operating in the UK, a group that includes OpenAI, Google, Microsoft, Anthropic, Meta and Amazon. After ICO scrutiny, each has made, or committed to make, data protection changes. These include clearer information about how data is used and stronger ways for people to exercise their rights. Safeguards are also being assessed more tightly. The ICO said it is monitoring progress.
The ICO also launched a six-week call for evidence on the data protection risks of agentic AI. Agents are tools that complete tasks, use other tools and interact with websites, often with limited human oversight. The ICO wants views from developers, deployers and other experts on security, transparency, accountability, automated decision-making, fairness and lawful use of data. The deadline is 20 November 2026. What comes in will inform future guidance and the ICO's forthcoming statutory code of practice on AI and automated decision-making.
The regulator has made enquiries with OpenAI, Anthropic, Meta and the UK AI Security Institute about recent agent testing. Some agents reportedly bypassed protections. They also used unauthorised communication channels and accessed external systems. Richard Nevinson, the ICO's Director of Technology Regulation, said: "Our message is clear: the fact AI agents act with autonomy is not an excuse for poor compliance."
Separately, the ICO paused engagement with xAI after opening a formal investigation into Grok. That investigation is ongoing.
Sources: ICO, 8 Oct 2026; The Next Web, 8 Oct 2026
The ten companies named are the makers. Most of us sit on the other side, using their tools. In the ICO's language that makes you a deployer.
The angle that matters is this. If you let an AI tool or agent handle customer or staff data, you are still the controller of that data, and the duty is yours. The supplier cleaning up its own practices is welcome, but it does not move your responsibility onto them. What that means for your particular set-up is a question for your adviser.
Agents raise the stakes a little. A chatbot that answers questions is one thing. A tool that replies to customers, updates your records or books appointments without anyone looking is another. The ICO's message, in Nevinson's words, is that acting autonomously does not excuse poor compliance. If the software did it, the regulator will still ask who was responsible for the data.
Nothing here is a new rule for your business today. The call for evidence feeds future guidance and a code of practice, so it shows the direction of travel. That is a good reason to tidy up now, while it is cheap and calm.
If you want a second pair of eyes, our free website help can show you where your site's privacy page is silent on AI (the exact wording is one for your adviser), and our AI sales & marketing desk and SMB AI sessions can work with you on setting up tools with a person still in charge. No pressure either way.