BuildNot yet confirmed elsewhere1 publisher2 min readPublished
MI5 head Ken McCallum announces an AI Threat Assessment Centre
MI5 director general Ken McCallum announced a new AI Threat Assessment Centre and said AI already significantly amplifies some threats, especially in cyber. He set the warning inside a claim that MI5 now commits more resources to state threats than it did in the Cold War.
The Engineer · Build desk
What happened
- McCallum warned that without proper safeguards the technology could raise the capabilities of malicious actors with a relatively low level of training.
- He said AI could also help create destructive biological weapons, although that has not yet happened.
- The number of cases linked to state threats has quadrupled over the last decade, according to McCallum.
- Russia's spying capability weakened after more than 1,000 officers were expelled from abroad after 2022, McCallum said, and Russia now pays criminals to carry out its orders.
Why it matters
- constraint The AI statements describe a direction of risk. A defender cannot turn that into a control change or a test case.
- decision The centre's first published assessment will decide whether it is an input to threat models or only a heading in a speech.
- precedent AI now sits in the annual speech beside Russia, Iran, China and terrorism. We'd expect later speeches to be judged on whether AI gets casework counts like the ones the state-threat figures have.
Only one of McCallum's AI statements is in the present tense, and it arrives without a case attached. The nearest cyber example in the speech is the attack on Polish energy infrastructure. He cited it as part of Russia's escalating hybrid war against European countries [8]. The report does not describe the new centre's remit, staffing or budget, and does not say whether AI played any part in the Polish attack. The other two AI statements are forward-looking by their own terms. One depends on safeguards being absent [3]. The other concerns something that has not happened [4].
The account here is mezha.net's relay of CNN's report [12], and the quotations in this piece are translated from its Ukrainian text.
An assessment centre, going by its name, produces assessments. In our view the test is whether MI5 publishes them in a form operators can use, with incidents and counts attached.
McCallum framed the period as one in which the stability of the post-Cold War years gave way to an "era of greater confrontation" [1]. He said the heightened threat level is not a "temporary deviation" [11]. Every count in the report belongs to casework. Counter-terrorism and state-threat missions are now roughly equal in scale, he said [7]. Iran-linked plots led to more than 50 arrests, mostly on suspicion of planning attacks on Jewish communities [10].
The low-skill warning sits naturally beside the Russia passage. A state that has lost officers and now pays criminals for its tasks gains most from tools that make less-trained people effective. That link is our inference. McCallum, as reported, did not make it.
What to watch
- Whether MI5 or the government publishes the centre's remit, staffing and budget.
- Whether the centre's first assessment includes incidents or counts comparable to MI5's casework figures.
- MI5's own text of the speech, to check the translated quotations against the original wording.
Clarity's read
What the record supports and how the coverage leans. The claims behind it follow.
Reality
- Evidence35
- Adoption
- Insufficient
- Hype gap
- Insufficient
- Incentives
- Insufficient
- Confidence50
Claim ledger
Ranked by verification strength, evidence, and original report placement.
- [1]
In MI5's annual threat speech on Tuesday, Ken McCallum said the stability of the post-Cold War years had been replaced by an "era of greater confrontation".
ReportedSupportedSource: mezha.net, relaying CNN; quotation translated from UkrainianView cited source - [2]
McCallum said AI "already significantly amplifies" some threats, especially in cyber, as he announced the creation of a new AI Threat Assessment Centre.
- [3]
McCallum warned that without proper "safeguards" the technology could increase the capabilities of "malicious actors with a relatively low level of training".
ReportedSupportedSource: mezha.net, relaying CNN; quotation translated from UkrainianView cited source - [4]
McCallum said AI could also help create destructive biological weapons, although this has not yet happened.
- [5]
McCallum said that in 2026 MI5 directed more resources at state threats than it did during the Cold War.
- [6]
According to McCallum, the number of cases linked to state threats has quadrupled over the last decade.
- [7]
McCallum said MI5's operational missions against terrorism and against state threats are now roughly equal in scale.
- [8]
McCallum called Russia the 'number one aggressor' and pointed to an escalation of its hybrid war against European countries, citing a cyberattack on Polish energy infrastructure and an attempt to use an explosive-laden drone to destroy a cargo plane in Germany.
- [9]
McCallum said Russia's spying capability was significantly weakened after more than 1,000 of its officers abroad were expelled following the 2022 full-scale invasion of Ukraine, and that Russia instead began to "pay criminals to carry out its orders".
ReportedSupportedSource: mezha.net, relaying CNN; quotation translated from UkrainianView cited source - [10]
McCallum said more than 50 people have been arrested over Iran-linked plots, mostly on suspicion of planning attacks on Jewish communities in the UK.
- [11]
McCallum said the heightened level of threat is not a "temporary deviation".
ReportedSupportedSource: mezha.net, relaying CNN; quotation translated from UkrainianView cited source - [12]
mezha.net's report of McCallum's speech is based on reporting by CNN.
Sources
1 independent publisher whose own reporting we read for this story.
Topics and entities
Follow any of these and your For You feed starts watching them — no settings page required.