Breach of Confidence: 24 April 2026
嗯,用户让我用中文总结一下这篇文章的内容,控制在一百个字以内,而且不需要以“文章内容总结”或“这篇文章”这样的开头,直接写描述即可。好的,我先仔细看看这篇文章的内容。 这篇文章看起来是每周的安全博客摘要,涵盖了多个安全相关的主题。首先提到了ChatGPT无法可靠地进行事实核查,这可能影响依赖它的业务策略。然后是一个关于荷兰和比利时边境的秋千的有趣故事,引出了互联网忽视国界但后来又试图重建国界的讨论。 接下来是开源领导力与企业领导力的对比,开源维护者通过愿景吸引志愿者,而商学院则教授威胁员工的方法。Meta公司监控员工的每一个动作来训练AI模型,同时裁员,这两件事被坚称无关。IT帮助台被假冒的情况增多,利用合法软件进行攻击。 还有一个24岁的年轻人通过短信诈骗窃取了800万美元,整个安全行业都没能阻止这种简单却有效的方法。朝鲜利用AI进行大规模加密货币盗窃,通过虚假工作机会进行社会工程学攻击。网络访问权限在暗网上以15英镑的价格出售,显示了网络犯罪模式的转变。 最后提到AI暴露了互联网免费服务背后的成本问题,广告补贴模式难以为继。整篇文章充满了对技术、安全和互联网未来的深刻洞察和讽刺。 现在需要把这些内容浓缩到100字以内。我得抓住每个部分的核心点:ChatGPT的问题、秋千引出的边界讨论、开源领导力、Meta监控裁员、假冒IT帮助台、短信诈骗、朝鲜AI攻击、网络访问价值低、AI成本问题。 可能需要舍弃一些细节,比如具体案例的数据或具体人物的故事。重点放在现象和趋势上。例如:ChatGPT的事实核查问题影响业务;秋千象征边界消失;开源领导力优于传统管理;Meta监控与裁员;IT帮助台被滥用;短信诈骗成功;朝鲜利用AI诈骗开发者;网络访问低价出售;AI暴露互联网成本问题。 这样组合起来应该能在100字左右涵盖所有要点。 </think> 文章探讨了ChatGPT的事实核查问题、国际边界消失的象征意义、开源领导力的优势、Meta监控员工与裁员事件、假冒IT帮助台攻击、短信诈骗案例、朝鲜利用AI进行加密货币盗窃以及网络访问权限的价值等话题,并指出AI暴露了互联网免费服务背后的成本问题。 2026-4-24 09:3:36 Author: securityboulevard.com(查看原文) 阅读量:13 收藏

The post Breach of Confidence: 24 April 2026 appeared first on Javvad Malik.

I spent an hour this week explaining to someone that no, ChatGPT cannot reliably fact-check itself, and yes, that’s a problem when your entire business strategy depends on it being right. They looked at me like I’d just told them Father Christmas works part-time at Argos.

The Swing That Crosses Borders 40 Times a Minute

There’s a swing somewhere between the Netherlands and Belgium that crosses an international border every time it moves. A Belarusian father watched his kids play on it and said he hoped borders would one day be unnecessary. I can’t stop thinking about that whilst watching governments panic about encrypted messaging apps and data sovereignty. We built the internet to ignore borders, then spent thirty years trying to put them back.

https://europeancorrespondent.com/en/r/borders-are-a-construct-but-this-swing-isnt

Open Source Leaders Are Better Managers Than Your Actual Manager

Open source maintainers have no authority, no budget, and no ability to fire anyone. All they can do is articulate a vision clearly enough that strangers volunteer their evenings to make it real. Meanwhile, business schools charge £30,000 to teach leadership models that boil down to “threaten people until they comply.” If you want to learn how to actually lead, stop reading Harvard Business Review and start watching how Linus Torvalds (ok maybe not the best example) gets people to care about kernel patches… but you git what I mean (see what I did there)

https://allthingsopen.org/articles/power-vs-influence-open-source-leadership

Meta’s Watching Every Keystroke Whilst Sacking Thousands

Meta’s now logging every single keystroke and mouse click from employees to train AI models. They’re also laying off thousands of people. These two things are definitely unrelated, they insist. Just a normal Tuesday at the office. The surveillance is to improve productivity. The redundancies are to improve margins. The fact that they’re happening simultaneously is merely a coincidence wrapped in a dystopia dressed as innovation.

https://www.bbc.co.uk/news/articles/cvglyklz49jo

Your IT Helpdesk Is Probably a Bloke in Lagos

Microsoft Teams is being used to impersonate your actual IT helpdesk, and tools like Quick Assist are doing exactly what they’re designed to do: let someone remote into your machine. The attack works because it’s not an attack. It’s legitimate software doing legitimate things for illegitimate people. You can’t patch trust.

https://www.bleepingcomputer.com/news/security/microsoft-teams-increasingly-abused-in-helpdesk-impersonation-attacks/

A 24-Year-Old from Dundee Stole $8 Million via Text Message

A lad from Dundee (I assume he’s a good wee lad, never got into any trouble) just pleaded guilty to stealing $8 million through SMS phishing. He’s facing twenty years in prison. He didn’t exploit a zero-day. He didn’t write custom malware. He just sent convincing text messages and people believed him. The entire security industry exists to prevent this exact thing, and yet here we are, watching someone barely old enough to rent a car retire early on the proceeds of typos and urgency.

https://krebsonsecurity.com/2026/04/scattered-spider-member-tylerb-pleads-guilty/

North Korea’s Using AI to Steal Crypto at Industrial Scale

North Korea’s moved on from banks. Now they’re targeting developers with fake job offers, using AI to make the scam scalable. They’re not hacking systems. They’re hacking hope. You apply for a role, go through interviews, get sent a test project, and boom. Your wallet’s empty. It’s social engineering industrialised, and it’s working because developers are tired, underpaid, and desperate for remote work that doesn’t involve legacy PHP codebases.

https://expel.com/blog/inside-lazarus-how-north-korea-uses-ai-to-industrialize-attacks-on-developers/

Your Network Access Is Worth £15

Initial access brokers are selling your corporate network credentials for about fifteen quid on dark web forums. Not because you’re unimportant. Because stealing credentials at scale is now so efficient that the market treats them like Tesco Value burgers (Aldi ones maybe?). The entire business model of cybercrime has shifted from precision heists to bulk commodity trading. Your breach isn’t personal. It’s just inventory.

https://www.computing.co.uk/feature/2026/cybercrime-who-are-the-initial-access-brokers

AI Broke the Internet’s Business Model

We spent a decade assuming compute was basically free. Turns out it’s not. AI just exposed that lie. Every search query, every chatbot response, every generated image costs real money in real datacentres burning real electricity. The entire internet ran on ads subsidising free services. Now those services cost actual money to run, and nobody’s quite sure who’s supposed to pay for it. Spoiler: it’s you. It’s always you.

https://stratechery.com/2026/mythos-muse-and-the-opportunity-cost-of-compute

___

That’s it for this week. If any of this made you laugh, wince, or forward it to someone who needs to hear it, reply and tell me. Or find me on Bluesky where I’m usually complaining about vendor pitches or tea temperatures.

Stay cynical.

*** This is a Security Bloggers Network syndicated blog from Javvad Malik authored by j4vv4d. Read the original post at: https://javvadmalik.com/2026/04/24/breach-of-confidence-24-april-2026/


文章来源: https://securityboulevard.com/2026/04/breach-of-confidence-24-april-2026/
如有侵权请联系:admin#unsafe.sh