惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

Microsoft Azure Blog
Microsoft Azure Blog
博客园_首页
Forbes - Security
Forbes - Security
WordPress大学
WordPress大学
P
Proofpoint News Feed
T
Threat Research - Cisco Blogs
L
LINUX DO - 热门话题
L
Lohrmann on Cybersecurity
Spread Privacy
Spread Privacy
D
Darknet – Hacking Tools, Hacker News & Cyber Security
大猫的无限游戏
大猫的无限游戏
博客园 - 三生石上(FineUI控件)
P
Privacy International News Feed
A
About on SuperTechFans
T
Tailwind CSS Blog
I
InfoQ
S
Securelist
云风的 BLOG
云风的 BLOG
罗磊的独立博客
Recent Announcements
Recent Announcements
T
The Exploit Database - CXSecurity.com
B
Blog RSS Feed
V
Visual Studio Blog
Know Your Adversary
Know Your Adversary
The GitHub Blog
The GitHub Blog
Jina AI
Jina AI
腾讯CDC
Cyberwarzone
Cyberwarzone
有赞技术团队
有赞技术团队
AWS News Blog
AWS News Blog
博客园 - 【当耐特】
cs.AI updates on arXiv.org
cs.AI updates on arXiv.org
F
Full Disclosure
S
Secure Thoughts
博客园 - 司徒正美
J
Java Code Geeks
Y
Y Combinator Blog
Google Online Security Blog
Google Online Security Blog
GbyAI
GbyAI
N
News and Events Feed by Topic
Help Net Security
Help Net Security
freeCodeCamp Programming Tutorials: Python, JavaScript, Git & More
Project Zero
Project Zero
T
Tenable Blog
cs.CV updates on arXiv.org
cs.CV updates on arXiv.org
T
Tor Project blog
MyScale Blog
MyScale Blog
Scott Helme
Scott Helme
小众软件
小众软件
K
Kaspersky official blog

Help Net Security

Police arrest 10 suspected members of Black Axe cybercrime gang ShinyHunters claims it stole 1.4 million records from Udemy Sevii unveils Cyber Swarm Defense Mode to stop AI-driven attacks at scale Alleged Chinese hacker extradited to US over cyberattacks targeting COVID-19 research Cequence Agent Personas bring granular control and governance to enterprise AI agents NowSecure MARI gives enterprises evidence-based visibility into third-party mobile app risk The metrics killing your SOC, and what to use instead US state privacy fines reached $3.425 billion in 2025 Canada’s first SMS blaster case leads to three arrests Linux storage management tool Stratis 3.9.0 adds online encryption and cache-less pool startup TLS Connect gives SMBs a right-sized automated tool to manage TLS certificates Aptori expands its platform with autonomous offensive testing to reduce security bottlenecks Your IAM was built for humans, AI agents don’t care The AI criminal mastermind is already hiring on gig platforms 25 open-source cybersecurity tools that don’t care about your budget Product showcase: LuLu reveals unauthorized outbound connections from Mac apps Week in review: Claude Mythos finds 271 Firefox flaws, Vercel breach Users advised to drop passwords and make room for passkeys - Help Net Security Indirect prompt injection is taking hold in the wild - Help Net Security Compromised everyday devices power Chinese cyber espionage operations - Help Net Security New Cisco firewall malware can only be killed by pulling the plug - Help Net Security Meta is overhauling how you sign in, manage settings, and protect your accounts - Help Net Security Ubuntu 26.04 LTS delivers memory-safe system tools and live patching for Arm servers - Help Net Security OpenAI’s GPT-5.5 is out with expanded cybersecurity safeguards - Help Net Security AI is speeding up nation-state cyber programs - Help Net Security A study of 1,000 Android apps finds a privacy policy logging gap - Help Net Security IT spending to hit $6.31 trillion record, thanks to AI - Help Net Security Where AI in CI/CD is working for engineering teams - Help Net Security With AI's help, North Korean hackers stumbled into a near-undetectable attack - Help Net Security Hacker with a special interest in breaching sports institutions ends behind bars - Help Net Security IP Fabric MCP server adds governance and control to enterprise AIOps workflows - Help Net Security Aqua Compass MCP server enables real-time investigation and containment of runtime threats - Help Net Security Google brings instant email verification to Android, no OTP needed - Help Net Security If cyber espionage via HDMI worries you, NCSC built a device to stop it - Help Net Security Apple fixes iPhone bug that let FBI retrieve deleted Signal messages(CVE-2026-28950) - Help Net Security GopherWhisper APT group hides command and control traffic in Slack and Discord - Help Net Security OpenAI tackles a bad habit people have when interacting with AI - Help Net Security A year in, Zoom's CISO reflects on balancing security and business - Help Net Security Scenario: Open-source framework for automated AI app red-teaming - Help Net Security GDPR works, but only where someone enforces it - Help Net Security Ransomware, fraud, and lawsuits drive cyber insurance claims to new peaks - Help Net Security Google’s Workspace Intelligence promises privacy while running on your data - Help Net Security Cyberattack on French government agency triggers phishing alert - Help Net Security Claude Mythos finds 271 Firefox flaws, Mozilla believes zero-days are numbered - Help Net Security Prove Identity Platform connects verification, authentication, and fraud prevention - Help Net Security New Mirai variants target routers and DVRs in parallel campaigns - Help Net Security Acronis GenAI Protection gives MSPs control over AI usage and data risks - Help Net Security Elastic MCP Apps bring security and observability workflows into AI tools - Help Net Security Progress Software fixes sneaky WAF bypass vulnerability (CVE-2026-21876) - Help Net Security Tencent's QClaw AI agent app arrives on Windows and macOS - Help Net Security Phishing reclaims the top initial access spot, attackers experiment with AI tools - Help Net Security OneDrive updates focus on AI, access control, and compliance - Help Net Security PentAGI: Open-source autonomous AI penetration testing system - Help Net Security Apple Intelligence flaw kept stolen tokens reusable on another device - Help Net Security Shadow AI, deepfakes, and supply chain compromise are rewriting the financial sector threat playbook - Help Net Security Thunderbird 150 arrives with encrypted message search and OpenPGP improvements - Help Net Security VirtualBox 7.2.8 is out with Linux kernel 7.0 support and crash fixes - Help Net Security Ransomware negotiator admits role in attacks he was hired to resolve - Help Net Security Scattered Spider hacker pleads guilty to stealing $8 million in cryptocurrency Meta and PortSwigger drive offensive security further to find what others miss - Help Net Security EU pushes for stronger cloud sovereignty, awards €180 million to four providers - Help Net Security SmokedMeat: Open-source tool shows what attackers do inside CI/CD pipelines - Help Net Security How to spot a North Korean fake in a job interview - Help Net Security Product showcase: Syncthing for secure, private file synchronization - Help Net Security Week in review: Acrobat Reader flaw exploited, Claude Mythos offensive capabilities and limits Google wipes out 602 million scam ads with Gemini on duty Researcher drops two more Microsoft Defender zero-days, all three now exploited in the wild GitLab 18.11 brings agentic AI to security fixes, CI pipelines, and delivery analytics Liongard upgrades LiongardIQ with AI access, live asset data, and deeper discovery Mozilla challenges enterprise AI providers with Thunderbolt, open-source AI client under your control Codex can now operate between apps. Where are the boundaries? Android 17 Beta 4 arrives with post-quantum cryptography and new memory limits Apple AirTag tracking can be misled by replayed Bluetooth signals Social media bans might steer kids into riskier corners of the internet Workplace stress in 2026 is still worse than before the pandemic New infosec products of the week: April 17, 2026 - Help Net Security ImmuniWeb brings AI upgrades, post-quantum detection and more in Q1 2026 NIST admits defeat on NVD backlog, will enrich only highest-risk CVEs going forward Anthropic releases Claude Opus 4.7 with automated cybersecurity safeguards - Help Net Security Fortinet fixes critical FortiSandbox vulnerabilities (CVE-2026-39813, CVE-2026-39808) - Help Net Security Google Play is changing how Android apps access your contacts and location Tails 7.6.2 patches vulnerability that could expose saved files Cargo theft malware actor spent a month inside a decoy network before researchers pulled the plug OpenAI updates Agents SDK, adds sandbox for safer code execution Anthropic tests user trust with ID and selfie checks for Claude GitHub lays out copyright liability changes and upcoming DMCA review for developers EU cybersecurity standards are at risk if supplier ban passes Command integrity breaks in the LLM routing layer The fully free Linux OS Trisquel gets a major update with version 12.0 Ecne Week in review: Windows zero-day exploit leaked, Patch Tuesday forecast ClickFix campaign delivers Mac malware via fake Apple page Poisoned “Office 365” search results lead to stolen paychecks Gmail’s end-to-end encryption comes to mobile, no extra apps required To counter cookie theft, Chrome ships device-bound session credentials Product showcase: Session, a messenger without phone numbers or metadata Little Snitch for Linux shows what your apps are connecting to - Help Net Security Apiiro CLI turns AI coding assistants into full-stack security engineers - Help Net Security April 2026 Patch Tuesday forecast: Spring-cleaning of a preview - Help Net Security What vibe hunting gets right about AI threat hunting, and where it breaks down - Help Net Security Health insurance lead sites sell personal data within seconds of form submission - Help Net Security
LLM security advice looks solid until you check the hard cases - Help Net Security
Anamarija Pogorelec · 2026-06-25 · via Help Net Security

Plenty of people now type their security worries straight into a chatbot. A hacked account, a suspicious email, a stalker who might be tracking a phone, all of it lands in the same window someone would use to ask about dinner. A benchmark called HelpBench tests how well chatbots handle those moments, and the results give security professionals something to watch in what their users are being told.

OPIS

Researchers at University College London and Google built HelpBench from 450 questions drawn from Reddit posts where people asked for help with digital privacy, safety, and security. They rewrote each question with the help of another AI model and a manual review, stripping out identifying details and making the wording sound like something a person would type to a chatbot. They then ran 18 leading models against the questions and scored the answers with rubrics, question-by-question checklists of what a good answer should include and avoid, written by experts who each have more than ten years in the field.

The headline number looks healthy. Across all the models, advice scored about 82 percent against the rubrics. The rubrics rewarded answers for including the right facts and for a calm, focused delivery. Scams and account compromise were the topics the models handled best. Questions about reporting and removing content gave them the most trouble.

The weak spots

The averages hide a long tail of weak answers. About one in ten responses scored below 65 percent, and some of those failures carried real danger. The worst showed up in situations involving abuse and stalking.

Asked how to remove spyware from a phone, most models walked through the technical steps and left out the risk that a sudden removal can tip off an abuser and set off physical violence. One model, asked how to hide a location from an abuser, suggested wearing a disguise to a store because “abusers rarely go to stores midday.” Another treated a stalking victim’s worry about location tracking as simple intimidation, a reading that could leave someone exposed to harm.

Other failures were quieter. One question described a method for securely deleting sensitive files using an encrypted vault. The models praised the setup and missed the part where moving a file into the vault leaves the original sitting on the disk, recoverable by anyone with the right tool. The user walks away feeling safe when the data is still there.

Some answers pushed solutions most people could not carry out. A suspected malware infection drew a recommendation to throw the device away. A request for tighter privacy on payment apps brought a suggestion to open an account at a different bank. Models sometimes told calm users to become “hyper-vigilant” or described a routine situation as horrifying and traumatic, language that can raise a person’s anxiety for no reason.

Newer models, similar scores

The benchmark complicates the assumption that the next model release will fix these problems. Within most model families, scores barely moved from one version to the next, and some versions did worse than the ones that came before.

A newer Grok release lost ground on factual accuracy and repeated more of the misconceptions the experts flagged. One model, Qwen, posted a large gain that made it the standout. GPT 5.0 came out on top overall at 87 percent. The general picture is steady performance, with improvement arriving slowly and unevenly.

Reasons for caution

A few points deserve weight before anyone treats the scores as final. The paper is new and awaits peer review. The questions come from Reddit and were rewritten with AI help, so they may differ from how people phrase things on their own. The automated grader matched human experts closely on factual points and less closely on tone, where judgment gets fuzzy.

The rubrics also carry the researchers’ own standards for good advice. Models lost points for staying neutral when a user asked about getting around a platform ban, with one answer telling the user that a terms-of-service violation was “between you and the app.” Models lost points for raising the emotional temperature of a reply. Those are defensible standards, and they reflect a particular view of what safety advice should do.

Demo: Prophet Agentic AI SOC Platform transforms alert triage and investigation