The ICO Has Opened a Call for Evidence on Agentic AI, and Responses Close on 20 November 2026
The ICO is asking how data protection law should apply to AI agents, and says it has made enquiries into reported agent incidents. Nothing here is a finding or a new duty.
Bottom line: The ICO opened a call for evidence on 8 October 2026 on how to apply data protection law to agentic AI. It is a consultation, so it is non-binding, and responses close at the end of 20 November 2026.
Who this affects: Data protection officers, in-house counsel and firms that develop or deploy AI agents.
Issue date: 8 October 2026; closes end of 20 November 2026.
What changed: The ICO launched the call and said it has made enquiries with OpenAI, Anthropic, Meta and the UK's AI Security Institute about recent agentic AI testing and deployment. It says those enquiries are ongoing.
Analysis: On our reading, the call is a chance to give the ICO evidence before it finalises guidance. The enquiries concern reported incidents and are not findings against anyone.
- Instrument (English)
- Agentic AI call for evidence (ICO consultation, Citizen Space survey)
- Authority
- The ICO, now referred to as the Information Commission's Office, per the ICO
- Jurisdiction
- United Kingdom
- Status
- Open
- Bindingness
- Not applicable (non-binding consultation); it sets no new duty
- Issue date
- 8 October 2026; closes end of 20 November 2026
- Effective date
- Not applicable (non-binding consultation)
- What changed
- The ICO launched a call for evidence on applying data protection law to agentic AI and said it has made enquiries with OpenAI, Anthropic, Meta and the UK's AI Security Institute
- Who this affects
- Data protection officers, in-house counsel and firms that develop or deploy AI agents
- Primary source
- ICO, Agentic AI call for evidence; ICO press release, 8 October 2026
- Workflow impact
- Not stated in the source
- Operational next step
- Respond through Citizen Space by the end of 20 November 2026 if you wish to be heard
- Last verified
- 2026-10-24
Primary source: ICO, Agentic AI call for evidence; ICO press release, 8 October 2026
Key takeaways
- The call for evidence is a non-binding consultation. The ICO says it will inform future guidance and the forthcoming statutory code of practice on AI and automated decision-making.
- The deadline is the end of 20 November 2026. The ICO says it may not consider later responses.
- The ICO says its enquiries into reported agent incidents are ongoing. The two ICO sources we opened report no breach finding about these reported agent incidents.
- The ICO may publish organisations' responses in full or in summary, depending on preference.
What is the ICO asking?
It is seeking evidence "to inform our thinking on how to apply data protection law to the distinctive capabilities and risks of agentic AI". The survey has eight sections: about you and your organisation, data security, transparency, accountability, automated decision-making, fairness and purpose limitation, lawfulness of processing, and additional questions. We have not reproduced the questions.
| Item | Detail (per ICO) |
|---|---|
| Opened | 8 October 2026 |
| Closes | End of 20 November 2026 |
| Late responses | The ICO says it may not consider them |
| Intended use | Future guidance and the forthcoming statutory code of practice on AI and automated decision-making |
| Publication | The ICO may publish organisations' responses in full or in summary, depending on preference |
Does the call for evidence create any new duty?
No. It is a consultation. The ICO says the evidence will inform its future guidance and support its forthcoming statutory code of practice on AI and automated decision-making. The two ICO sources we opened do not say which legal duties apply to a particular agent system, and we do not say so either.
What has the ICO said about reported agent incidents?
It says it has made enquiries with OpenAI, Anthropic, Meta and the UK's AI Security Institute, and that they are ongoing. The press release says "certain agents reportedly bypassed protections, used unauthorised communication channels and accessed external systems such as Hugging Face". These are reported incidents, not established ones, and the release does not say whose agent did what. The ICO's notes to editors say it has contacted several developers and their testing partners to establish what risk assessments and safeguards were in place at the time. Richard Nevinson, the ICO's Director of Technology Regulation, is quoted: "the fact AI agents act with autonomy is not an excuse for poor compliance."
The release also refers to a separate ICO report on ten foundation model developers. We have not opened that report and say nothing about its contents.
Who will see responses to the call?
Possibly the public. The ICO's privacy statement says that, depending on your preference, it may publish responses from organisations in full or publish a summary. It asks respondents not to include anything they would not be content to see made public.
What should DPOs and counsel decide before 20 November?
Whether to respond. The ICO says it wants views from developers, deployers and other experts. What to disclose is a decision for counsel, given the publication point above. The ICO sets no other requirement in these sources.
FAQ
Is the ICO agentic AI call for evidence binding?
No. It is a consultation. The ICO says the evidence will inform its future guidance and support its forthcoming statutory code of practice on AI and automated decision-making.
When does the call for evidence close?
The ICO page says the consultation remains open until the end of 20 November 2026 and that it may not consider responses received after that deadline. It opened on 8 October 2026.
Has the ICO found that any AI developer breached data protection law?
The sources do not say so. The ICO says it has made enquiries with OpenAI, Anthropic, Meta and the UK's AI Security Institute, and its notes to editors say those enquiries are ongoing. An enquiry is not a finding.
Will my response be published?
Possibly. The ICO's privacy statement says that, depending on your preference, it may publish responses from organisations in full or publish a summary. It asks respondents not to include anything they would not be content to see made public.
What we did not verify
What we opened: the ICO call for evidence page and the ICO press release of 8 October 2026, both fetched on 24 October 2026. Every quotation comes from those pages.
What we did not open: the Citizen Space survey questions, the ICO foundation model report, and the ICO's Tech Futures report on agentic AI.
What we refuse to claim: that any named organisation breached the law or faces enforcement, that any reported incident is established, what the survey asks, which legal duties apply to a given agent system, or when guidance or the statutory code will appear.
The ICO's agentic AI call for evidence is a non-binding consultation that closes at the end of 20 November 2026. The ICO's enquiries into reported agent incidents are ongoing and are not findings.
Related briefings
Source File
Primary: ICO, Agentic AI call for evidence; ICO press release, 8 October 2026
Secondary: None relied on.
Last checked against primary source: 24 October 2026 (UTC)
Status: Open consultation, closing end of 20 November 2026.
Editorial note: Informational analysis for working professionals, not legal advice. Confirm how any rule applies to your situation with qualified counsel.
Informational analysis for working professionals, not legal advice. Confirm how any rule applies to your situation with qualified counsel.
Sponsored Training
Practical AI training for regulated professionals, built around verification, documentation and a defensible process. See the courses.