L

Lord Tarassenko (CB)

Speaking in the House of Lords on 26 October 2026

Debate

Cyber Security and Resilience (Network and Information Systems) Bill

Contribution

My Lords, I declare my interest as senior adviser to the Alan Turing Institute. I wish to speak in support of Amendment 41, in the name of the noble Baroness, Lady Harding, to which I have added my name, and then briefly about my own Amendment 66 in this group. We heard during Committee not only about the out-of-control behaviour from agents running AI models from OpenAI, Google and Meta but about the unacceptable delays before the incidents were reported. Contrast that with our own AI Security Institute. When it found unsanctioned agent behaviour during its testing this summer—and this was confirmed to us by Henry de Zoete at a meeting in Portcullis House today—it immediately declared a security incident and contained it within an hour of the alert being raised, with all relevant machines isolated within that time. That is the standard we should be expecting. With Amendment 41, the developer must notify the affected organisation and its regulator within 24 hours. The duty applies where the developer is aware, or ought reasonably to be aware, so no company can avoid it by not looking. The report must say whether the model acted autonomously or was used by someone else. On 12 October, the Government said that they would look at strengthening this Bill so that serious AI incidents are reported. Their Amendment 70 gives the Secretary of State powers to investigate where control of an AI agent is lost. However, no one can investigate an incident they have not been told about. Amendment 41 is what makes Amendment 70 work. I would therefore be interested to find out whether the noble Baroness will be minded to accept Amendment 41. I turn briefly to Amendment 66 in my name, to which the noble Lords, Lord Clement-Jones and Lord Holmes of Richmond, and the noble Baroness, Lady Harding, have added their names. This amendment would require the AI Security Institute to audit frontier AI products before they were used in the network and information systems that run our hospitals, our energy networks and our water supply. Developers would submit their evaluation results using methods and thresholds set by AISI, with evidence of a tested means of suspending the system in the UK. Where the institute found a serious risk that was not mitigated, the Secretary of State would have to decide within 14 days whether to act and give reasons. As I have already said, I welcome the Government’s Amendment 70, which accepts that loss of control of an AI agent is a risk to essential services that should be addressed in this Bill. However, it operates after the event and does not mention the AI Security Institute. I would be grateful, therefore, if the Minister could tell the House what role the Government see for the institute—AISI—in assessing frontier models before they reach our critical systems and whether the power in Clause 29(1) could give it that role.

More from Lord Tarassenko (CB)

Other recent Hansard contributions by the same speaker.

About Hansard

Hansard is the official verbatim record of proceedings in the UK Parliament. Every word spoken in the Commons and Lords is recorded and published — this page is a single contribution from that record.