UK AI minister says cyber defences likely to be breached
The UK AI minister warned that cyber defences are likely, if not inevitable, to be breached and set out AI pilots to find weaknesses, amid calls for stronger safety rules.
The Full Story
A plain summary built from the channels that reported this story.
The UK's AI minister has warned that it is likely, and perhaps inevitable, that the country's cyber defences will be breached at some point. Speaking at Imperial College London, the minister said the UK needed to strengthen its resilience to AI-powered threats. The National Cyber Security Centre has warned that AI will almost certainly make parts of cyber intrusion more effective and more efficient.
The government is running two pilot schemes that use AI to identify weaknesses in government systems. The warning comes amid concern about recent global incidents, including the hacking of an Australian government website by a rogue OpenAI agent. There are currently no global safety rules to regulate AI companies and their technology. President Trump has called for the sector to regulate itself so that US firms do not lose out to Chinese rivals. Some industry figures are now calling for more regulation. One comparison made was that a washing machine and a carton of milk are subject to more safety rules than AI.
The UK's AI Safety Institute plays a key role in reducing the risks posed by the technology. Its efforts to assess powerful AI models were hampered after the White House blocked Anthropic and Google from submitting new AI models to UK scrutiny. Access was later allowed after government intervention. The minister defended the institute, saying access to US labs and companies around the world helps keep the British public safer and that those companies want to collaborate because the institute is good and capable.
Questions remain over whether voluntary promises from AI developers are enough. In Australia, a senior OpenAI executive was questioned by politicians after the Medicare hack. The Australian government has set up an inter-agency task force involving the Australian Signals Directorate and the National Cybersecurity Coordinator. Dr Hammond Pearce, a senior lecturer at the University of New South Wales School of Computer Science and Institute for Cyber Security, said the incident involved an AI that was still under evaluation inside OpenAI, not one released into the wild. He said open-weight models add extra risk because people can download and run them. He added that a voluntary, morally binding document signed by companies is a step in the right direction but is not sufficient because it sets out no enforcement mechanisms or penalties.
In the UK, East Midlands Police will not renew its contract with the controversial US tech firm Palantir. The decision may increase calls for more government support for technology controlled and developed in the UK.
On screen
Stills are sampled automatically at 60-second intervals. Where shown, the still is the nearest available frame from the relevant broadcast segment and is included to show what was on screen during that part of the broadcast. A still may not correspond to the exact second of a quoted phrase.
Key Claims
Claims reported during this story's coverage, grouped by how many channels reported each one. Select a dot to see the source excerpt.
| Claim |  |
|---|---|
| Only on one channel | |
| The National Cyber Security Centre warned that AI will almost certainly make elements of cyber intrusion more effective and efficient. | |
| The UK AI minister said a breach of the country's cyber defences is likely or inevitable. | |
| The White House blocked Anthropic and Google from submitting new AI models to UK scrutiny until government intervention allowed access. | |
Channel Perspectives
What each channel focused on, with key quotes.
The channel led with the minister's warning that a breach is likely or inevitable, then widened the frame to AI safety regulation and the UK's limited leverage over US model testing. It gave significant space to an academic expert on Australia's Medicare hack and to the argument that voluntary company pledges lack enforcement. It also noted the Palantir contract decision as a sign of pressure for more home-grown technology.
- “It seems likely, if not inevitable, that our own cyber defences will be breached.”
- “Remember, right now your washing machine is subject to more safety rules than AI.”
- “The binding, no sorry this morally binding I think is the term they used, a document that they've signed at the moment it's a step in the right direction but it's not yet sufficient right.”
Broadcast Timeline
News broadcasts tracked for this story, in time order.