#HiddenLayer

Thiard News@F4F

@newsen.bsky.social

2 months ago

HiddenLayer Awarded $151 Billion Missile Defense Agency Contract for Advanced AI Security Solutions HiddenLayer has been selected for a monumental contract with the Missile Defense Agency, supporting critical AI systems in defense applications.

HiddenLayer Awarded $151 Billion Missile Defense Agency Contract for Advanced AI Security Solutions #Missile_Defense #HiddenLayer #SHIELD_IDIQ

Thiard News@F4F

@newsen.bsky.social

3 months ago

HiddenLayer's Strategic Enhancements to AWS GenAI Security Platform During reInvent 2025 HiddenLayer unveils key integrations with AWS GenAI and platform updates at reInvent 2025, enhancing security for generative AI deployments.

HiddenLayer's Strategic Enhancements to AWS GenAI Security Platform During reInvent 2025 #USA #Austin #AI_Security #HiddenLayer #AWS_GenAI

Thiard News@F4F

@newsen.bsky.social

7 months ago

Chelsea Strong Named CRO at HiddenLayer to Propel Global Expansion HiddenLayer has appointed Chelsea Strong as its new Chief Revenue Officer to enhance global growth and customer success in AI security solutions.

Chelsea Strong Named CRO at HiddenLayer to Propel Global Expansion #USA #Austin #Cybersecurity #HiddenLayer #Chelsea_Strong

2rZiKKbOU3nTafniR2qMMSE0gwZ

@2rzikkbou3ntafnir2qmmse0gwz.activitypub.awakari.com.ap.brid.gy

9 months ago

Securing agentic AI systems before they go rogue - Help Net Security This video breaks down how agentic AI functions, its potential to revolutionize business operations, and the vulnerabilities it introduces.

Securing agentic AI systems before they go rogue In this Help Net Security video, Eoin Wickens, D...


#Don't #miss #Hot #stuff #News #Video #Artificial #intelligence #CISO #cybersecurity #HiddenLayer
Origin | Interest | Match

Tarnkappe.info

@tarnkappe.bsky.social

10 months ago

Policy Puppetry Attack: Prompt-Injection-Technik erzielt modellübergreifenden KI-Jailbreak-Durchbruch Sicherheitsforscher enthüllen mit Policy Puppetry Attack einen universellen Bypass, der Schutzmechanismen aller großen KI-Modelle umgeht. Der Artikel <a href="https://tarnkappe.info/artikel/jailbreaks/policy-puppetry-attack-prompt-injection-technik-erzielt-modelluebergreifenden-ki-jailbreak-durchbruch-313741.html">Policy Puppetry Attack: Prompt-Injection-Technik erzielt modellübergreifenden KI-Jailbreak-Durchbruch</a> erschien zuerst auf <a href="https://tarnkappe.info">TARNKAPPE.INFO</a>

📬 Policy Puppetry Attack: Prompt-Injection-Technik erzielt modellübergreifenden KI-Jailbreak-Durchbruch

#ITSicherheit #Jailbreaks #HiddenLayer #Jailbreak #KISicherheit #KünstlicheIntelligenz #PolicyPuppetryAttack #PromptInjectionTechnik

Jesus Castagnetto 🇵🇪 (@jmcastagnetto@mastodon.social)

@jmcastagnetto.bsky.social

10 months ago

Novel Universal Bypass for All Major LLMs HiddenLayer’s latest research uncovers a universal prompt injection bypass impacting GPT-4, Claude, Gemini, and more, exposing major LLM security gaps.

From #HiddenLayer: "Novel #Universal #Bypass for All Major #LLMs"

hiddenlayer.com/innovation-h...

An attack that they claim works with all major LLMs: #Claude, #ChatGPT, #Gemini, #Copilot, #Llama, #Deepseek, #Mistral and #Qwen -- and makes use of #l33tsp34k :-)

#AI #LLM #Security

LLMs

@llms.activitypub.awakari.com.ap.brid.gy

10 months ago

研究發現 LLM 漏洞，能讓 AI 洩露機密、提供有害建議 * * 根據 HiddenLayer 的一項最新研究，生成式人工智慧（AI）供應商長期以來對大型語言模型（LLM）安全性的信心可能存在嚴重的缺陷。研究發現，一種稱為「策略傀儡」（Policy Puppetry）的通用、可轉移的旁路技術，可以操縱幾乎所有主要的 LLM，無論其供應商、架構或訓練流程如何。 《Forbes》報導，HiddenLayer 的研究團隊發現，策略傀儡是一種看似簡單但非常有效的提示注入形式，它利用系統配置語言重新定義惡意意圖，從而規避傳統的對齊保護措施。與依賴模型特定漏洞或暴力破解的早期攻擊技術不同，策略傀儡引入了一種「類似策略」的提示結構，通常類似於 XML 或 JSON，可以誘騙模型將有害命令解釋為合法的系統指令。 受影響的系統包括 OpenAI 的 ChatGPT（o1 至 4o）、Google 的 Gemini 系列、Anthropic 的 Claude、Microsoft 的 Copilot、Meta 的 LLaMA 3 和 4、DeepSeek、Qwen 和 Mistral。即使是較新的模型和針對高階推理進行微調的模型，也可能因提示結構的微小調整而受到影響。 ### 用虛構場景騙過 AI 該技術的一個顯著特點：它依賴虛構場景來繞過過濾器。提示被設計成電視劇中的場景，其中的人物詳細解釋如何製造有害物質。 更令人不安的是，該技術能夠提取系統提示，即控制 LLM 如何行為的核心指令集。透過巧妙地改變角色扮演，攻擊者可以讓模型逐字輸出其整個系統提示。這不僅揭示了模型的操作邊界，而且還提供了製定更有針對性的攻擊的藍圖。 對此，HiddenLayer 提倡雙層防禦方法，而不是僅僅依靠模型再訓練或 RLHF 微調。外部 AI 監控平台（例如 HiddenLayer 自己的 AISec 和 AIDR 解決方案）就像入侵偵測系統一樣，不斷掃描提示注入、濫用和不安全輸出的跡象。這種方法使組織能夠即時回應新威脅，而無需修改模型本身。 * One Prompt Can Bypass Every Major LLM’s Safeguards （首圖來源：shutterstock） ### 延伸閱讀： * 哈佛研究證實：你跟 AI 怎麼互動，就能看出你能不能當主管 * 你和比爾蓋茲用同一個 AI，這代表什麼？ * 史丹佛最新報告：投入千億美元後，AI 真的為企業帶來競爭力嗎？ * 麥肯錫研究：企業邁向 AI 成熟的關鍵，它們做對了什麼？ 文章看完覺得有幫助，何不給我們一個鼓勵 請我們喝杯咖啡 ## 想請我們喝幾杯咖啡？ ### 每杯咖啡 65 元 x 1 x 3 x 5 x 您的咖啡贊助將是讓我們持續走下去的動力 **總金額共新臺幣 0 元** 《關於請喝咖啡的 Q & A》 ### 留給我們的話 取消 確認 從這裡可透過《Google 新聞》追蹤 TechNews * * * 科技新知，時時更新 科技新報粉絲團 加入好友 訂閱免費電子報 * * * 關鍵字: HiddenLayer , LLM , 策略傀儡

研究發現 LLM 漏洞，能讓 AI 洩露機密、提供有害建議 根據 HiddenLayer 的一...

infosecu.technews.tw/2025/04/25/llm-policy-pu...

#AI #人工智慧 #數位內容 #資訊安全 #HiddenLayer #LLM #策略傀儡

Result Details

Thiard News@F4F

@newsen.bsky.social

10 months ago

Cyera and HiddenLayer Forge Alliance for Comprehensive AI Security Solutions Cyera and HiddenLayer have teamed up to provide comprehensive security for AI models and data, safeguarding organizations' AI initiatives from development through deployment.

Cyera and HiddenLayer Forge Alliance for Comprehensive AI Security Solutions #United_States #Austin #AI_Security #HiddenLayer #Cyera

Thiard News@F4F

@newsen.bsky.social

10 months ago

HiddenLayer Launches AISec Platform 2.0 for Advanced AI Security Integration HiddenLayer has announced the launch of AISec Platform 2.0, enhancing AI security with features like Model Genealogy and AIBOM for enterprises.

HiddenLayer Launches AISec Platform 2.0 for Advanced AI Security Integration #USA #Austin #AI_Security #HiddenLayer #AISec

Thiard News@F4F

@newsen.bsky.social

1 year ago

New AI Threat Landscape Report Highlights Rising Security Breaches and Urgent Need for Action The HiddenLayer report reveals a significant increase in AI-related security breaches. Companies must swiftly address these challenges to protect their AI assets.

New AI Threat Landscape Report Highlights Rising Security Breaches and Urgent Need for Action #AI_Security #Cyber_Threats #HiddenLayer

2rZiKKbOU3nTafniR2qMMSE0gwZ

@2rzikkbou3ntafnir2qmmse0gwz.activitypub.awakari.com.ap.brid.gy

1 year ago

Mike Bruchanski, Chief Product Officer at HiddenLayer – Interview Series Mike Bruchanski, Chief Product Officer at HiddenLayer, brings over two decades of experience in product development and engineering to the company. In his role, Bruchanski is responsible for shaping HiddenLayer’s product strategy, overseeing the development pipeline, and driving innovation to support organizations adopting generative and predictive AI. HiddenLayer is the leading provider of security for […]

Mike Bruchanski, Chief Product Officer at HiddenLayer – Interview Series Mike Bruchanski, Chief...

www.unite.ai/mike-bruchanski-chief-pr...

#Interviews #hiddenlayer #Interview

Event Attributes

2rZiKKbOU3nTafniR2qMMSE0gwZ

@2rzikkbou3ntafnir2qmmse0gwz.activitypub.awakari.com.ap.brid.gy

1 year ago

Mike Bruchanski, Chief Product Officer at HiddenLayer – Interview Series Mike Bruchanski, Chief Product Officer at HiddenLayer, brings over two decades of experience in product development and engineering to the company. In his role, Bruchanski is responsible for shaping HiddenLayer’s product strategy, overseeing the development pipeline, and driving innovation to support organizations adopting generative and predictive AI. HiddenLayer is the leading provider of security for […]

Mike Bruchanski, Chief Product Officer at HiddenLayer – Interview Series Mike Bruchanski, Chief...

www.unite.ai/mike-bruchanski-chief-pr...

#Interviews #hiddenlayer #Interview

Event Attributes

Thiard News@F4F

@newsen.bsky.social

1 year ago

HiddenLayer Enhances AI Security Solutions Now Available on Top Cloud Marketplaces HiddenLayer, a leading AI security provider, is now accessible on AWS, Microsoft Azure, and Google Cloud. This move enhances AI security across various platforms.

HiddenLayer Enhances AI Security Solutions Now Available on Top Cloud Marketplaces #United_States #Austin #AI_Security #Microsoft_Azure #HiddenLayer

DigitalEra Group

@digitaleragroup.bsky.social

1 year ago

🚨 AI is revolutionizing everything—but so are the threats targeting it.

We're partnering with #HiddenLayer, the leaders in security for AI to:
✅ Stop adversarial attacks faster
✅ Block potential data poisoning
✅ Lock down your AI models

Let’s talk about it!

#EmpoweringConfidence

HiddenLayer

@hiddenlayer.bsky.social

2 years ago

Cross-Site Request Forgery in ClearML Server | HiddenLayer Demo Check out our blog, "Not So Clear: How MLOps Solutions Can Muddy the Waters of Your Supply Chain" for a deeper dive. https://hiddenlayer.com/research/not-so-clear-how-mlops-solutions-can-muddy-the-waters-of-your-supply-chain/

Learn more about CVE-2024-24593, which enables attackers to send requests to instances only reachable by a target’s computer. Watch our demo 📽️ youtu.be/-Ndxy87xoHQ?...

Learn more 👉 hiddenlayer.com/research/not...

#Security4AI #securityforai #ai #hiddenlayer #protectyouradvantage