UK's Ministry of Defence pins hopes on AI to stop the next massive email blunder
- Reference: 1754480712
- News link: https://www.theregister.co.uk/2025/08/06/mod_taps_aussie_ai_shop/
- Source link:
Aussie startup Castlepoint Systems announced this morning that the MoD had selected it to provide what it calls AI-powered data control.
The security shop promises that its explainable AI tech automates the control over complex datasets and helps reduce the likelihood of human error leading to serious leaks.
[1]
Rachael Greaves, CEO at Castlepoint Systems, said: "The MoD faces a complex challenge in managing vast and sensitive datasets in the knowledge that even a single case of data leak or loss can be catastrophic. I'm pleased that after undertaking a very thorough global search, Castlepoint was selected by the MoD as the best solution to solve this problem.
[2]
[3]
"Castlepoint, with explainable AI and true autoclassification at its core, can increase labeling accuracy and coverage without disrupting the essential work of MoD personnel. We are a trusted technology provider for public-sector organisations and enterprises in Australia and New Zealand, and having now established our global headquarters in London, we look forward to delivering our proven solutions to many more organisations in the UK."
The company claims its technology is already used by two-thirds of government departments in Australia, and has contracts in New Zealand too.
Afghan data leak
The deal follows fresh details regarding the [4]MoD's infamous 2021 data leak where it mistakenly publicized the identities of nearly 19,000 Afghans who worked with British forces during the conflict with the Taliban.
In what is considered one of the most, if not the most, damaging data breaches in UK history, the people's names were exposed thanks to a classic CC-not-BCC email blunder from the UK's Afghan Relocations and Assistance Policy (ARAP) unit.
[5]
The Taliban vowed to punish anyone who helped the British during the war, meaning exposing that data potentially threatened the lives of thousands.
After a super-injunction was lifted in July, it was revealed that in addition to the Afghans' identities, those of around 100 British officials, including SAS troops and MI6 spies, were also exposed.
The [6]BBC also reported that one individual obtained a copy of the exposed identities and published a snippet to Facebook, threatening to leak more.
[7]
The individual, who has not been named, was reportedly one of the Afghans who had their resettlement application rejected by the UK. Their application was reconsidered on an expedited basis following the threats, and it is understood they are now in the UK.
Is AI the answer?
The buzz around AI and its potential applications for security defenders has been building for years, and although many organizations are deploying the latest tech, few seem to know how to configure it securely.
[8]War Games: MoD asks soldiers with 1337 skillz to compete in esports
[9]Britain's billion-pound F-35s not quite ready for, well, anything
[10]UK eyes new laws as cable sabotage blurs line between war and peace
[11]UK CyberEM Command to spearhead new era of armed conflict
That notion was on display at the country's National Cyber Security Centre's (NCSC) annual [12]CYBERUK conference earlier this year. Peter Garraghan, CEO at Mindgard and professor of distributed systems at Lancaster University, asked a crowd filled with infosec pros to raise their hands if they fully understood the security risks associated with AI system controls, and not a single one was raised out of the 200-strong crowd.
"So everyone's using generative AI, but no one has a grasp of how secure it is in the system," Garraghan replied. "The cat's out of the bag."
The NCSC launched a report shortly before Garraghan's session, warning organizations against rushing AI deployments due to the increased attack surface these systems present.
At the same time, it said failing to [13]AI-ify cyber defenses could lead to them becoming significantly more vulnerable to evolving, AI-empowered security threats by 2027.
An NCSC spokesperson told The Register at the time: "Organizations and systems that do not keep pace with AI-enabled threats risk becoming points of further fragility within supply chains, due to their increased potential exposure to vulnerabilities and subsequent exploitation. This will intensify the overall threat to the UK's digital infrastructure and supply chains across the economy.
"The NCSC's supply chain guidance is designed to help organizations gain effective control and oversight over their supply chains. We encourage organizations to use this resource to better understand and manage the risks.
"This is also why market incentives need to exist, to drive up resilience at scale, at an increased velocity." ®
Get our [14]Tech Resources
[1] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/publicsector&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=2&c=2aJN8CNVLpITvPuNhV1BGXAAAAFc&t=ct%3Dns%26unitnum%3D2%26raptor%3Dcondor%26pos%3Dtop%26test%3D0
[2] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/publicsector&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44aJN8CNVLpITvPuNhV1BGXAAAAFc&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0
[3] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/publicsector&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33aJN8CNVLpITvPuNhV1BGXAAAAFc&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0
[4] https://www.theregister.com/2021/09/21/mod_email_fail_afghan_interpreters_data/
[5] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/publicsector&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44aJN8CNVLpITvPuNhV1BGXAAAAFc&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0
[6] https://www.bbc.co.uk/news/articles/cj4ek9njknvo
[7] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/publicsector&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33aJN8CNVLpITvPuNhV1BGXAAAAFc&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0
[8] https://www.theregister.com/2025/07/29/mod_asks_soldiers_with_1337/
[9] https://www.theregister.com/2025/07/15/uk_f35_failings/
[10] https://www.theregister.com/2025/07/02/uk_cable_sabotage_law/
[11] https://www.theregister.com/2025/06/04/uk_cyberem_command_details/
[12] https://www.theregister.com/2025/05/14/cyberuk_ai_deployment_risks/
[13] https://www.theregister.com/2025/07/30/firms_are_neglecting_ai_security/
[14] https://whitepapers.theregister.com/
It's the combination that does most damage!
Watch any course, any course you like, about AI Agents and they all say the same thing, the "human in the loop" is mandatory 'cos AI is simply not up to par or to be trusted. These are leading industry AI experts, but when did politicians ever listen to experts?!
AI solves everything
I've no problem with the company being Australian, but I note its most recent UK subsidiary accounts show the only assets as £10 in the bank, no employees, and two prior compulsory striking off actions by Companies House. Doesn't fill one with confidence.
Re: AI solves everything
They're probably Chinese!
It doesn't matter if your security threat is "AI-empowered" it's still not getting past my closed firewall ports.
What could go wrong.
https://cybernews.com/security/ai-tools-data-breaches-workplace-security-risks/
They aren't paying to make it secure.
They are buying a scapegoat. Next time it all goes pear-shaped, they will just blame the AI. That may be the real value of AI for corporations and government.
There are much cheaper solutions than buying AI software, but as they are spending tax payers' money, not their own, cost is not an issue. And AI is trending. Just as NFTs, blockchain and the metaverse once were. Nobody ever got sacked for jumping on a bandwagon.
Re: They aren't paying to make it secure.
The law needs to ensure that gap is well and truly plugged.
Join the Great Revolt
We must negate the machines-that-think. Humans must set their own guidelines. This is not something machines can do. Reasoning depends upon programming, not on hardware, and we are the ultimate program! Our Jihad is a "dump program." We dump the things which destroy us as humans!
I've come to the conclusion that the current "AI" isn't just bullshit, but dangerous bullshit that requires an organised resistance and a Butlerian solution. Despite being a huge fan of the late Iain M Banks' works.
Re: Join the Great Revolt
I don't want to see it all dumped but it's all getting a bit gung-ho as "AI" is grasped as a panacea.
Anyone else remember the stickers with the slogan "To er is human, to completely foul things up you need a computer"?