The UK government plans to use money set aside in the Defence Investment Plan to develop its own agentic artificial intelligence (AI) incident response capabilities to be put to work during major cyber incidents, the UK’s artificial intelligence (AI) minister, Kanishka Narayan, has said.
In an update delivered this week to Parliament on the fast-evolving abilities of frontier AI models, Narayan said: “we have committed £115 million to two new programmes: one on AI biosecurity, and one to build a UK Government agentic AI incident response capability.”
The proposals sit alongside a wider programme of work on agentic system security, including recently-published National Cyber Security Centre (NCSC) guidance on the subject, the development of the Cyber Shield defensive programme, and the Cyber Security and Resilience Bill, which is working its way through Parliament. This is complemented by the government Cyber Action Plan, which is backed by £210m of funding to address resilience in public services.
Following several of incidents over the past few weeks – most notably frontier model ‘escapes’ at OpenAI and the UK’s AI Security Institute (AISI) lab – Narayan said he was working closely with security minister Dan Jarvis to ensure learnings from these upsets inform the development of the UK’s future cyber framework.
“We will consider whether protections for increasingly autonomous AI systems should be clarified or strengthened through the Cyber Assessment Framework, the forthcoming statutory code of practice or NCSC technical guidance. AISI will provide evidence and technical expertise to support that work,” he said.
“As the UK invests in the capacity to use increasingly powerful AI, we will invest in parallel in the capacity to understand it, to control it, and to recover when something goes wrong,” added Narayan.
Agentic incident responders are not a new concept, with many security operations centres (SOCs) already actively deploying them to autonomously triage, investigate and remediate malicious cyber activity.
Often they are sold on the basis that if deployed appropriately, they should enable human cyber teams to focus their energies more efficiently, and avoid becoming overwhelmed by alerts and burning out.
Sukhveer Sanghera, co-founder of Potentially AI, said: “The government standing up an agentic-AI incident-response capability is a good step – we all saw what happened with the recent ‘breakouts.’ Importantly, these incidents were agents running in a
test environment with the guardrails off.
“What worries me more is a person who knows what they’re doing. They won’t ask one model for the whole thing. They’ll ask three or four for a piece each, and none of those questions will look like much. Each model checks its own conversation. Nobody is checking across them.
“Funding incident response is right. But you want to catch this earlier, and that means looking at how models are being used together,” he said.
Transnational risks
In his statement, Narayan also pledged to work with the UK’s foreign partners and allies to address the “inherently transnational” risks of AI.
“The systems involved are developed and deployed across borders, and no country can address them alone, as has been the case in wider technology for the last three decades,” he said.
“The UK will continue to work closely with international partners, including through the AISI’s relationships with counterpart bodies overseas and the NCSC’s peer agencies. We will approach this challenge with both confidence and urgency, ensuring that the UK can harness the benefits of frontier AI while managing the risks responsibly.”
